{ "updated": "2026-11-03", "models": [ { "slug": "drex-max-finance", "name": "Drex Max Finance", "provider": "Nace.AI", "providerId": "drex-max-finance", "discovered": "2026-11-03", "released": "2026-11-01", "verification": "Native catalog verified; account access varies", "specialized": "decision", "context": 262144, "maxOutput": null, "questionLimit": 512, "modalities": [ "text", "image" ], "tools": false, "sourceUrl": "https://drex.nace.ai/docs/drex/guides/drex-max", "officialUrl": "https://drex.nace.ai/docs/drex/reference/models", "apiDocumentationUrls": [ "https://drex.nace.ai/docs/drex/reference/limits", "https://drex.nace.ai/docs/drex/reference/pricing" ], "documentedRates": { "input": 0.03, "output": 0, "cached": null, "label": "Native Drex API \u00b7 USD" }, "note": "Check financial records against rules and answers you define\u2014for example, whether an invoice exceeds its approved purchase order.", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "capabilityNote": "Reads text and up to eight PNG, JPEG or WebP images. Returns yes/no probabilities, choices and ordered scores, not a written analysis. The finance specialization has no independent accuracy or calibration result in this comparison.", "contextNote": "The 262,144-token limit covers the whole request: shared text, images, all questions and formatting. Each added question consumes that budget. At most 512 questions, 255 options per question, 8 images and 4 MiB serialized input are accepted.", "pricingNote": "Costs $0.03 per million billed input tokens, including image tokens. Output and failed requests are not charged. Use the successful response\u2019s input usage to calculate the bill.", "billingNote": "Image tokens already belong to billed input. Output and failed requests are not charged; use successful response usage totals.", "accessNote": "Check GET /v1/models with the API key and environment you will use. Send drex-max-finance explicitly; availability can differ by account during rollout. It has its own serving capacity, with no fallback. Omitting the model still selects drex-v2.0.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "hydrogen-1-0", "name": "Hydrogen 1.0", "provider": "Deep Principle", "providerId": "deep-principle/Hydrogen-1.0", "sourceUrl": "https://huggingface.co/deep-principle/Hydrogen-1.0", "officialUrl": "https://huggingface.co/deep-principle/Hydrogen-1.0", "discovered": "2026-11-03", "released": null, "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": null, "note": "Explore scientific questions and draft research workflows for expert review. Access to the model files requires accepting the publisher\u2019s conditions.", "deployment": "checkpoint", "verification": "Gated checkpoint listed", "availability": "restricted", "specialized": "science", "checkpointRevision": "b71dd62f26fa459ca548117c11c6ad567af2faba", "contextNote": "The public card advertises 1M context and 128K output, subject to runtime resources and configuration. Exact file-level limits could not be verified without accepting the repository\u2019s access conditions; the 4,096-token example is not the maximum.", "capabilityNote": "Post-trained from GLM-5.3 for chemistry, materials science and life-science questions. Published demonstrations use external tools and research environments. Neither the parent\u2019s scores nor the publisher\u2019s scientific results are independent measurements for this profile.", "accessNote": "Hugging Face requires login and agreement to share account information before file access. The public card labels BF16 weights MIT and independently authored code/docs Apache 2.0; review the actual notices when obtaining access. Its serving example uses eight GPUs and custom code, and requires generation_config.json for correct answer stopping. Separate NVFP4/MXFP4 versions target different hardware.", "pricingNote": "No verified token tariff or measured local operating cost. The publisher also advertises NeuroSpark API and Mira access; those routes are not established by the Hugging Face inference-provider list.", "apiDocumentationUrls": [ "https://blog.deepprinciple.com/introducing-hydrogen-1.0/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "longcat-2-5-gateway", "name": "LongCat 2.5 \u00b7 AI Gateway", "provider": "Meituan", "providerId": "meituan/longcat-2.5", "sourceUrl": "https://vercel.com/ai-gateway/models/longcat-2.5", "officialUrl": "https://vercel.com/ai-gateway/models/longcat-2.5", "discovered": "2026-11-03", "released": "2026-11-02", "context": 1048576, "maxOutput": 131072, "modalities": [ "text", "image" ], "tools": true, "note": "Try LongCat 2.5 through an existing Gateway account, with the limits and billing of that hosted offer.", "verification": "Hosting offer verified", "identityLabel": "AI Gateway offer", "documentedRates": { "input": 0.3, "output": 1.2, "cached": 0.006, "label": "Vercel AI Gateway \u00b7 USD" }, "capabilityNote": "Gateway advertises tools, structured output and thinking controls at none, low, medium and high. These are supported controls, not measured quality differences or a verified mapping to native thinking budgets.", "contextNote": "Gateway declares 1,048,576 combined tokens and a 131,072 output ceiling. Keep these hosted limits separate from the conflicting native output documentation.", "accessNote": "The stable hosted offer replaces the preview entry in Gateway\u2019s current catalog. This is a hosting option for the same base model, not another model release.", "pricingNote": "Gateway currently lists $0.30 input, $0.006 cached input and $1.20 output per million tokens. Check its quote before budgeting beyond the native provider\u2019s temporary promotion.", "apiDocumentationUrls": [ "https://vercel.com/changelog/longcat-2-5-now-available-on-ai-gateway" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "longcat-2-5", "name": "LongCat 2.5", "provider": "Meituan", "providerId": "LongCat-2.5", "sourceUrl": "https://longcat.ai/platform/docs/", "officialUrl": "https://longcat.ai/platform/docs/", "discovered": "2026-11-03", "released": "2026-11-02", "context": 1048576, "maxOutput": null, "modalities": [ "text", "image" ], "tools": true, "note": "Use text and pictures in coding and multi-step agent work. Check the limits of your chosen service before sending a large project.", "announced": "2026-11-01", "verification": "Native API verified", "documentedRates": { "input": 0.3, "output": 1.2, "cached": 0.006, "label": "LongCat native promotion \u00b7 USD" }, "contextNote": "The native API shares 1,048,576 tokens between input and output. Its parameter reference allows 262,144 output tokens, while the quick-start guide still says 128K. Confirm the usable output ceiling before relying on a long answer.", "capabilityNote": "Thinking can be enabled or disabled. The parameter guide permits a thinking budget up to 262,144 tokens but does not establish an omitted default. Temperature is fixed by thinking mode; other supplied values are overridden. No matched independent effort or speed comparison is available here.", "accessNote": "Create an API key and fund the account. The API changelog dates availability November 2; the announcement is dated November 1. Compatible chat and agent interfaces are documented. Public downloadable 2.5 weights are not verified.", "pricingNote": "Limited-time native rates are $0.30 input, $0.006 cached input and $1.20 output per million tokens. The promotion has no verified end date. The platform bill is authoritative; failed requests are not billed. The pricing page\u2019s 500K-output worked example is not a supported output limit.", "apiDocumentationUrls": [ "https://longcat.ai/platform/docs/api/model-params", "https://longcat.ai/platform/docs/api-pay-as-you-go", "https://longcat.ai/platform/docs/ChangeLog.html", "https://longcat.ai/blog/longcat-2.5" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "giotto-solo", "name": "Giotto Solo", "provider": "Giotto", "providerId": null, "sourceUrl": "https://www.giotto.ai/", "officialUrl": "https://www.giotto.ai/", "discovered": "2026-11-02", "released": null, "context": null, "maxOutput": null, "modalities": [], "tools": null, "note": "Consider Giotto\u2019s smaller reasoning model when you need a private deployment on your own compatible hardware.", "verification": "Provider deployment offer documented", "availability": "restricted", "capabilityNote": "Giotto positions Solo for single-GPU operation. This is a provider claim, not a verified memory requirement or a measured comparison with Alto.", "accessNote": "Private deployment and appliance options lead to sales contact. A public weight repository, exact Solo API ID and self-serve access terms are not verified. The Alto API alias and tariff are not assigned to Solo.", "contextNote": "Supported request limits and input formats need confirmation for the deployment you choose.", "pricingNote": "Request a quote for the model, hardware and support you need. No public Solo token tariff is verified.", "apiDocumentationUrls": [ "https://www.giotto.ai/offering", "https://www.giotto.ai/contact-us?interest=download" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "giotto-alto", "name": "Giotto Alto", "provider": "Giotto", "providerId": "giotto-model", "sourceUrl": "https://www.giotto.ai/api", "officialUrl": "https://www.giotto.ai/api", "discovered": "2026-11-02", "released": null, "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": null, "note": "Use Giotto\u2019s hosted reasoning model for document and coding workflows, or discuss a private deployment with its team.", "verification": "Native API and pricing documented", "documentedRates": { "input": 2.47, "output": 12.35, "cached": 0.25, "write": 2.47, "label": "Giotto website USD display" }, "capabilityNote": "The documented Responses interface accepts text and offers medium or high thinking effort; the client defaults to medium. Event streaming reports completed workflow events rather than each text token. Independent quality and effort comparisons are pending.", "contextNote": "An exact public context window and answer-length limit are not established.", "accessNote": "The API page advertises Alto and uses the request alias giotto-model. This is not a pinned version identifier. Access requires account activation and a credential for the selected subscription or token-credit product. Private GPU deployment is arranged through sales; no public weight download is verified.", "pricingNote": "The website\u2019s USD selector shows $2.47 input, $0.25 cache reads, $12.35 output and $2.47 cache writes per million tokens. Its default EUR display is \u20ac2.10/\u20ac0.21/\u20ac10.50/\u20ac2.10 respectively. Confirm the checkout currency and applicable plan; do not treat display conversions as a guaranteed future exchange rate.", "apiDocumentationUrls": [ "https://giottoai.zendesk.com/hc/en-001/articles/37073730796818-API", "https://www.giotto.ai/offering" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "sensenova-loop-9b-preview", "name": "SenseNova Loop 9B Preview", "provider": "SenseNova", "providerId": "sensenova/SenseNova-Loop-9B-Preview", "sourceUrl": "https://huggingface.co/sensenova/SenseNova-Loop-9B-Preview", "officialUrl": "https://huggingface.co/sensenova/SenseNova-Loop-9B-Preview", "discovered": "2026-11-02", "released": "2026-10-30", "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": null, "note": "Run a downloadable text-and-image model for document questions, coding and tool-assisted work. Its unusual design needs compatible serving software.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "eda5d06eec244f9bb078fd193255fc615720a1f9", "capabilityNote": "The model reuses some of its layers twice while producing an answer. This is a fixed part of its design, separate from the optional thinking toggle. Publisher benchmark results do not establish independent quality or speed here.", "contextNote": "The card describes training sequences up to 262,144 tokens, and the configuration uses that position limit. A universally supported serving limit and maximum answer length are not established. The 256-token example is not a model ceiling.", "accessNote": "Apache 2.0 project license and verified weight shards plus custom model code. The documented environment uses Transformers 5.18.0, PyTorch 2.11.0 and CUDA 13.0. SGLang and vLLM instructions require the specified implementation branches; a generic install command is not enough. Check memory for your actual input length.", "pricingNote": "No verified hosted token tariff or local operating-cost measurement. Include hardware and software maintenance when comparing it with an online API.", "apiDocumentationUrls": [ "https://github.com/OpenSenseNova/SenseNova-Loop", "https://github.com/OpenSenseNova/SenseNova-Loop/blob/main/LICENSE" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gpt-6-astra-law", "name": "GPT-6 Astra for Law", "provider": "OpenAI", "providerId": "gpt-6-astra-law", "sourceUrl": "https://developers.openai.com/api/docs/pricing", "officialUrl": "https://developers.openai.com/api/docs/pricing", "discovered": "2026-11-02", "released": null, "context": null, "maxOutput": null, "modalities": [], "tools": null, "note": "A legal-work configuration now listed by OpenAI. Confirm access and try it on documents you can check before planning a rollout.", "verification": "Pricing verified; access details pending", "specialized": "legal", "documentedRates": { "input": 10, "output": 50, "cached": 1, "label": "OpenAI promotional API rates \u00b7 USD" }, "capabilityNote": "OpenAI lists this as a legal model configuration. Supported input formats, tools, thinking settings and task results are not yet established in the reviewed public documentation.", "contextNote": "Context and output limits are not verified for this exact configuration. Ordinary Astra limits are not assumed to apply.", "accessNote": "The exact configuration ID appears on the official pricing page. A public model-detail page and general availability terms are not verified; confirm access with OpenAI.", "pricingNote": "Published rates per million tokens are $10 input, $1 cached input and $50 output. The promotion is available at least through December 31, 2026; this does not establish a January 1 increase or a later price. Cache-write and long-context rules for this configuration remain unverified.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "ling-3-1-flash-1019-gateway", "name": "Ling 3.1 Flash 1019 \u00b7 AI Gateway", "provider": "InclusionAI", "providerId": "inclusionai/ling-3.1-flash-1019", "sourceUrl": "https://vercel.com/ai-gateway/models/ling-3.1-flash-1019", "officialUrl": "https://vercel.com/ai-gateway/models/ling-3.1-flash-1019", "discovered": "2026-10-31", "released": "2026-10-30", "context": 1000000, "maxOutput": 262144, "modalities": [ "text" ], "tools": true, "note": "Use Ling 1019 through Gateway, but confirm billing before leaving an application running: its announced free period ends today.", "identityLabel": "AI Gateway offer", "verification": "Hosting offer verified", "documentedRates": { "input": 0, "output": 0, "cached": null, "label": "Temporary Gateway promotion \u00b7 USD" }, "capabilityNote": "Gateway lists text input, tool calls and optional thinking at none, low, medium or high. Check the thinking setting explicitly; the default is not documented. These controls are not measured effort-by-effort results.", "contextNote": "Gateway lists 1,000,000 tokens shared by input and output, with a 262,144 output ceiling. If you run the downloadable model yourself, check its separate limits; its card and configuration still disagree.", "accessNote": "A hosted offer for the existing 1019 revision, served by Novita. It is not another base-model release.", "pricingNote": "The free promotion is scheduled to end November 3, 2026. At this review the catalog still returns zero; no extension, exact cutoff timezone or paid rate is verified. This is a transitional quote, not a reliable ongoing cost estimate.", "apiDocumentationUrls": [ "https://vercel.com/changelog/ling-3-1-flash-1019-now-available-on-ai-gateway-for-free" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "clover-decision-preview", "name": "Clover decision model \u00b7 announced", "provider": "DiffuSpace", "providerId": null, "sourceUrl": "https://huggingface.co/blog/Diffuspace/clover-35b-a3b-preview", "officialUrl": "https://huggingface.co/blog/Diffuspace/clover-35b-a3b-preview", "discovered": "2026-10-31", "released": null, "context": null, "maxOutput": null, "modalities": [], "tools": false, "note": "A planned model for choosing from defined answers, such as routing a ticket or checking whether a condition is met.", "announced": "2026-10-30", "availability": "announced", "verification": "Coming soon", "specialized": "decision", "capabilityNote": "The publisher describes separate independent-question and joint-answer modes. In joint mode, answers can influence each other; this is different from evaluating every question independently.", "accessNote": "The decision model is promised for a later release. No exact download, API ID, launch date, price or supported capacity is verified. The Clover chat preview is a separate artifact.", "pricingNote": "No published price or measured operating cost. Publisher decision tests do not establish an independent ranking.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "clover-35b-a3b-preview", "name": "Clover 35B A3B Preview", "provider": "DiffuSpace", "providerId": "Diffuspace/Clover-35B-A3B-Preview", "sourceUrl": "https://huggingface.co/Diffuspace/Clover-35B-A3B-Preview", "officialUrl": "https://huggingface.co/Diffuspace/Clover-35B-A3B-Preview", "discovered": "2026-10-31", "released": "2026-10-30", "context": 131072, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Try a downloadable coding and reasoning model on your own hardware. This preview needs its supplied serving code and substantial memory.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "4b0402b65e523d0183a14c431955438714cb16ff", "capabilityNote": "A text-only diffusion model derived from Qwen3.6-35B-A3B, with 35 billion total parameters and about 3 billion active per step. It supports thinking and tool-call formatting. Publisher speed and coding results are not independent measurements here.", "contextNote": "The 131,072-token limit includes the prompt and generated answer. Generation reserves complete blocks, so leave room for the final block. The example output budget is not a verified maximum.", "accessNote": "Apache 2.0 weights with upstream attribution requirements. The supplied Transformers implementation uses custom repository code, Transformers 5.19.0 and PyTorch 2.10 or newer. BF16 weights alone are about 69 GB; running the model needs additional memory. The publisher\u2019s lighter demos do not establish a universal hardware requirement. Separate SGLang and llama.cpp setup guides are still promised.", "pricingNote": "No verified hosted token price. Compare hardware and serving costs using your actual workload.", "apiDocumentationUrls": [ "https://huggingface.co/blog/Diffuspace/clover-35b-a3b-preview", "https://github.com/DiffuSpace/Clover-35B-A3B-Preview", "https://huggingface.co/Diffuspace/Clover-35B-A3B-Preview/blob/main/LICENSE" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gpt-6-rosalind", "name": "GPT-6 Rosalind Research", "provider": "OpenAI", "providerId": "gpt-6-rosalind", "sourceUrl": "https://developers.openai.com/api/docs/changelog", "officialUrl": "https://developers.openai.com/api/docs/changelog", "discovered": "2026-10-31", "released": "2026-10-30", "context": null, "maxOutput": null, "modalities": [], "tools": null, "note": "Help approved life-sciences teams analyze evidence, plan experiments and work with scientific tools.", "availability": "restricted", "specialized": "research", "verification": "Restricted research access", "documentedRates": { "input": 15, "output": 75, "cached": 1.5, "label": "Restricted API, short context \u00b7 USD" }, "capabilityNote": "Specialized biological reasoning and scientific tool use. Public context limits, output limits, input formats and thinking settings are not verified.", "accessNote": "Requires Rosalind Research approval for the organization and user. Work and Codex access does not automatically include API access, which requires separate approval. Discovery approval is insufficient. Approved use is internal research; customers cannot expose this access to outside users.", "contextNote": "The pricing table distinguishes short and long context, but the reviewed sources do not establish this model\u2019s threshold or capacity. Confirm the limit for your approved account before sending large documents.", "pricingNote": "Short-context rates per million tokens: $15 input, $1.50 cache reads and $75 output. Long-context rates: $30/$3/$112.50. Cache writes cost $18.75 for 30 minutes or $37.50 for 12 hours at short context; long-context writes cost $37.50/$75. Confirm eligibility and the context boundary before estimating a large research job.", "apiDocumentationUrls": [ "https://developers.openai.com/api/docs/pricing", "https://help.openai.com/en/articles/20001193-gpt-rosalind-for-life-sciences-research" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "parsy", "name": "Parsy", "provider": "Aleph Alpha", "providerId": null, "sourceUrl": "https://aleph-alpha.com/en/blog/parsy-distilling-an-agentic-parsing-harness/", "officialUrl": "https://aleph-alpha.com/en/blog/parsy-distilling-an-agentic-parsing-harness/", "discovered": "2026-10-30", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Extract text, tables and page locations from difficult documents, so an assistant can point readers to the source passage.", "announced": "2026-10-29", "availability": "restricted", "verification": "Customer access documented", "capabilityNote": "A document-parsing model that returns content, reading order and bounding boxes. It is not a general chat assistant.", "accessNote": "Available through Aleph Alpha customer projects. No self-serve API ID, downloadable weights, public token limit or tariff is established.", "pricingNote": "Request a quote for your document workflow and test difficult scans, tables and page references before committing.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "ling-3-1-flash-base", "name": "Ling 3.1 Flash Base", "provider": "InclusionAI", "providerId": "inclusionAI/Ling-3.1-flash-base", "sourceUrl": "https://huggingface.co/inclusionAI/Ling-3.1-flash-base", "officialUrl": "https://huggingface.co/inclusionAI/Ling-3.1-flash-base", "discovered": "2026-10-30", "released": "2026-10-27", "context": 262144, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Use this pretrained foundation for further training and model research. It is not the ready-to-chat Flash 1019 model.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "c2cad79e46fcb7b46d96f9cf876ce1f2418b6e64", "capabilityNote": "A base checkpoint for training rather than a finished assistant. Do not transfer the instruction-tuned model\u2019s performance or tool behavior.", "contextNote": "The configuration sets 262,144 positions. This is not an independently tested long-document quality result.", "accessNote": "The model card labels these weights MIT. Check the license included with your chosen download before reuse.", "pricingNote": "Local hardware, training and serving costs depend on your setup; no hosted token tariff is assigned.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lumma-fev-0-6b-v2", "name": "Lumma-Fev 0.6B v2 \u00b7 announced", "provider": "FrontiersMind", "providerId": null, "sourceUrl": "https://huggingface.co/blog/FrontiersMind/lumma-fev-27b-v2", "officialUrl": "https://huggingface.co/blog/FrontiersMind/lumma-fev-27b-v2", "discovered": "2026-10-30", "released": null, "context": null, "maxOutput": null, "modalities": [], "tools": false, "note": "A smaller decision model is planned. Wait for the exact download and supported limits before choosing it.", "announced": "2026-10-29", "verification": "Coming soon", "availability": "announced", "specialized": "decision", "accessNote": "The publisher lists this size as coming soon. No verified download, launch date, capacity or benchmark result is available.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lumma-fev-0-1b-v2", "name": "Lumma-Fev 0.1B v2", "provider": "FrontiersMind", "providerId": "FrontiersMind/Lumma-fev-0.1B-v2", "sourceUrl": "https://huggingface.co/FrontiersMind/Lumma-fev-0.1B-v2", "officialUrl": "https://huggingface.co/FrontiersMind/Lumma-fev-0.1B-v2", "discovered": "2026-10-30", "released": null, "context": 2048, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Run simple ticket routing or topic classification locally with a small download. Test more demanding judgment tasks separately; small size does not imply comparable reasoning quality.", "announced": "2026-10-29", "verification": "Download verified", "availability": "available", "specialized": "decision", "accessNote": "Downloadable Apache 2.0 weights, about 300 MB, using lumma-fev 0.4 or later. The publisher supports CPU or GPU use. Repository creation is October 30; an exact public release time is not established.", "deployment": "checkpoint", "checkpointRevision": "f78136b6c3f8e0eb7b8abf46c511efb05df9ee3e", "capabilityNote": "A 150-million-parameter model derived from Nandi-Mini-150M. It returns yes/no and choice probabilities, not a written answer. Score questions are accepted but untested by the publisher.", "contextNote": "Shared text plus each question must fit in 2,048 tokens. Each question is evaluated independently. This model does not accept images.", "pricingNote": "No hosted token price or measured local operating cost is established. Publisher accuracy and calibration results remain separate from independent rankings.", "apiDocumentationUrls": [ "https://huggingface.co/FrontiersMind/Lumma-fev-0.1B-v2/blob/main/LICENSE", "https://huggingface.co/blog/FrontiersMind/lumma-fev-27b-v2" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lumma-fev-27b-v2", "name": "Lumma-Fev 27B v2", "provider": "FrontiersMind", "providerId": "FrontiersMind/Lumma-fev-27B-v2", "sourceUrl": "https://huggingface.co/FrontiersMind/Lumma-fev-27B-v2", "officialUrl": "https://huggingface.co/FrontiersMind/Lumma-fev-27B-v2", "discovered": "2026-10-30", "released": "2026-10-29", "context": 8192, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Run ticket routing, classification or yes/no checks on your own hardware, with probabilities for the answers you define.", "deployment": "checkpoint", "verification": "Download verified", "specialized": "decision", "questionLimit": null, "outputType": "Choice and yes/no probabilities; no generated reply.", "checkpointRevision": "85ee6576956ac972683d2bb987a79bfa7193e4d7", "capabilityNote": "Supports English, Hindi, Tamil, Kannada and Hinglish. Score questions are accepted but were not tested by the publisher. Image input requires the optional image runtime.", "contextNote": "The shared information plus one question must fit in 8,192 tokens. The parent model\u2019s larger window does not apply to this decision model.", "accessNote": "Apache 2.0 weights. Requires lumma-fev 0.3 or later; older versions reject this attention setup. The publisher suggests about 55 GB free GPU memory for 27B, or multiple GPUs. Allow extra room for your workload.", "pricingNote": "No hosted token price or measured local operating cost is established. Publisher accuracy and confidence tests remain separate from independent rankings.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lumma-fev-4b-v2", "name": "Lumma-Fev 4B v2", "provider": "FrontiersMind", "providerId": "FrontiersMind/Lumma-fev-4B-v2", "sourceUrl": "https://huggingface.co/FrontiersMind/Lumma-fev-4B-v2", "officialUrl": "https://huggingface.co/FrontiersMind/Lumma-fev-4B-v2", "discovered": "2026-10-30", "released": "2026-10-29", "context": 8192, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Run ticket routing, classification or yes/no checks on your own hardware, with probabilities for the answers you define.", "deployment": "checkpoint", "verification": "Download verified", "specialized": "decision", "questionLimit": null, "outputType": "Choice and yes/no probabilities; no generated reply.", "checkpointRevision": "a1477c170264ba50c950776c3e55ebd9df90df0b", "capabilityNote": "Supports English, Hindi, Tamil, Kannada and Hinglish. Score questions are accepted but were not tested by the publisher. Image input requires the optional image runtime.", "contextNote": "The shared information plus one question must fit in 8,192 tokens. The parent model\u2019s larger window does not apply to this decision model.", "accessNote": "Apache 2.0 weights. Requires lumma-fev 0.3 or later; older versions reject this attention setup. The publisher suggests about 10 GB GPU memory for 4B. Allow extra room for your workload.", "pricingNote": "No hosted token price or measured local operating cost is established. Publisher accuracy and confidence tests remain separate from independent rankings.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "drex-max-v1-1", "name": "Drex Max v1.1", "provider": "Nace.AI", "providerId": "drex-max-v1.1", "discovered": "2026-10-30", "released": "2026-10-28", "verification": "Native model verified", "specialized": "decision", "context": 262144, "maxOutput": null, "questionLimit": 512, "modalities": [ "text", "image" ], "tools": false, "sourceUrl": "https://drex.nace.ai/docs/drex/guides/drex-max", "officialUrl": "https://drex.nace.ai/docs/drex/reference/models", "apiDocumentationUrls": [ "https://drex.nace.ai/docs/drex/reference/limits", "https://drex.nace.ai/docs/drex/reference/pricing" ], "documentedRates": { "input": 0.03, "output": 0, "cached": null, "label": "Native Drex API \u00b7 USD" }, "note": "Answer up to 512 defined questions about text and images in one request. Use this version when the full set fits within one shared token budget.", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "capabilityNote": "The Max family accepts images. All questions see the same images and selected shared input. Use up to eight uploaded or inline PNG/JPEG/WebP images; ordinary URLs are treated as text. Images are resized, so fine detail can be lost. Uploaded images must belong to the API-key account and have no workspace; workspace files are refused.", "contextNote": "The 262,144-token limit covers the whole request: shared text, images, all questions and formatting. Each added question consumes that budget. At most 512 questions, 255 options per question, 8 images and 4 MiB serialized input are accepted.", "pricingNote": "Costs $0.03 per million billed input tokens, including image tokens. Output and failed requests are not charged. Use the successful response\u2019s input usage to calculate the bill.", "billingNote": "Image tokens already belong to billed input. Output and failed requests are not charged; use successful response usage totals.", "accessNote": "Send drex-max-v1.1 explicitly. It is separate from original drex-max; drex-latest and omitted model still select v2.0. No fallback to another model is documented for either Max version. Check the available-model list with your actual API key; availability can differ by environment.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "joey-1-gateway", "name": "Joey-1 \u00b7 AI Gateway", "provider": "Baseten / Base Labs", "providerId": "baseten/joey-1", "sourceUrl": "https://vercel.com/ai-gateway/models/joey-1", "officialUrl": "https://vercel.com/ai-gateway/models/joey-1", "discovered": "2026-10-30", "released": "2026-10-29", "context": 16384, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Route support tickets or rate text records against choices you define. This Gateway offer lists text input; use the native API if you need image input.", "specialized": "decision", "verification": "Hosting offer verified", "identityLabel": "AI Gateway offer", "accessNote": "This is a hosting offer for Joey-1, not another base model. Gateway lists 16,384 context tokens and no generated text.", "pricingNote": "Gateway lists $0.01 per million input tokens and zero output charge. Baseten native Flex and priority prices are separate.", "capabilityNote": "Gateway lists text input and choice, yes/no and rating questions. Native image support does not establish images on this offer.", "contextNote": "The shared Model API allows 16,384 tokens per question including the shared information. The local card also states a 16,384 trained sequence length. A larger dedicated-deployment claim is not the shared API limit.", "billingNote": "Use billed input from the API response, including question and image processing. No generated output-token charge.", "documentedRates": { "input": 0.01, "output": 0, "cached": null, "label": "AI Gateway \u00b7 USD" }, "apiDocumentationUrls": [ "https://www.baseten.co/library/joey-1/", "https://huggingface.co/baselabs/joey-1", "https://docs.baseten.co/reference/inference-api/system-one" ], "checkpointRevision": "9efbe485efbed9598dfca5294cbbe8da00e21908", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gemini-4-carbon", "name": "Gemini 4 Carbon", "provider": "Google", "providerId": null, "identityLabel": "Announced model", "announced": "2026-10-28", "released": null, "discovered": "2026-10-29", "checked": "2026-11-03", "verification": "Announced \u00b7 restricted access", "availability": "restricted", "context": null, "maxOutput": null, "modalities": [], "tools": null, "sourceUrl": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/", "officialUrl": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/", "apiDocumentationUrls": [ "https://deepmind.google/models/evals-methodology/gemini-4-carbon", "https://ai.google.dev/gemini-api/docs/models" ], "note": "Watch Carbon for demanding software-engineering and cybersecurity work. Access starts with selected defenders and trusted partners; most developers cannot switch to it yet.", "capabilityNote": "Google describes improvements in sustained coding work and multimodal understanding. Public input formats, tool support and API identifiers are not yet confirmed.", "accessNote": "Initial access is through Google\u2019s Fairwind program. Wider developer, business and consumer availability is planned, without a public launch date.", "contextNote": "Public context and output limits are not confirmed. A long-document benchmark does not establish the capacity your account will receive.", "dateNote": "Announced October 28; this is not a general-availability release.", "documentedRates": { "input": 2, "output": 10, "cached": 0.1, "label": "Announced introductory rates \u00b7 USD" }, "pricingNote": "Google announces $2 input and $10 output per million tokens, with cached input 95% cheaper ($0.10). After the introductory period, input/output become $4/$20. Carbon\u2019s promotion end date and other billing rules have not been announced.", "vendorBenchmarks": [ { "name": "SEC-bench Pro", "score": "88.2%", "meaning": "Finding software vulnerabilities", "note": "Google tested 183 JavaScript-engine tasks from the May 2026 set, running each task three times and averaging the single-attempt success rate.", "sourceUrl": "https://deepmind.google/models/evals-methodology/gemini-4-carbon" }, { "name": "ExploitGym", "score": "54.8%", "meaning": "Checking whether a vulnerability can be exploited", "note": "Google ran this test without internet access or deployment safeguards. Tests allowed six or ten hours per task; Google does not specify which limit produced this headline figure.", "sourceUrl": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/" } ], "vendorEvidenceNote": "Google published these results using its evaluation setup. They are not measurements from our matched independent ranking, and the tests differ from a real deployment. No effort-by-effort comparison or Agent Fit score is assigned.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "joey-1", "name": "Joey-1 \u00b7 native API", "provider": "Baseten / Base Labs", "providerId": "decision-models/baseten-joey-2b-latest", "sourceUrl": "https://docs.baseten.co/inference/model-apis/decision-models", "officialUrl": "https://docs.baseten.co/inference/model-apis/decision-models", "discovered": "2026-10-28", "released": "2026-10-29", "context": 16384, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Route support tickets, check a screenshot or rate a record against choices you define. Joey-1 is the small, low-price member of Base Labs\u2019 decision family.", "specialized": "decision", "verification": "Native API and download verified", "identityLabel": "Native API", "accessNote": "Available through Baseten and as a 2B download derived from Qwen3.5-2B. The card labels the downloadable weights MIT; check the license included with your chosen package. Use the supplied CUDA serving setup and review the package terms.", "pricingNote": "Native default input costs $0.01 per million tokens; output is free. Flex costs approximately 90% of default; priority defaults to twice the standard rate. These processing options do not establish measured speed gains.", "capabilityNote": "Returns yes/no probabilities, choices and ordered ratings. It does not write an explanation. The native API accepts text and images; test uncertain answers before allowing automatic actions.", "contextNote": "The shared Model API allows 16,384 tokens per question including the shared information. The local card also states a 16,384 trained sequence length. A larger dedicated-deployment claim is not the shared API limit.", "billingNote": "Use billed input from the API response, including question and image processing. No generated output-token charge.", "documentedRates": { "input": 0.01, "output": 0, "cached": null, "label": "Native Baseten API \u00b7 USD" }, "apiDocumentationUrls": [ "https://www.baseten.co/library/joey-1/", "https://huggingface.co/baselabs/joey-1", "https://docs.baseten.co/reference/inference-api/system-one" ], "checkpointRevision": "9efbe485efbed9598dfca5294cbbe8da00e21908", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "liquid-d1-omni-2b", "name": "Liquid d1 Omni 2B", "provider": "Liquid AI", "providerId": "LiquidAI/d1-omni-2B", "sourceUrl": "https://huggingface.co/LiquidAI/d1-omni-2B", "officialUrl": "https://huggingface.co/LiquidAI/d1-omni-2B", "discovered": "2026-10-28", "released": "2026-10-27", "context": 128000, "maxOutput": null, "modalities": [ "text", "image", "video", "audio" ], "tools": false, "note": "Classify a voice note or ask spoken questions about images and records, without generating a written reply.", "deployment": "checkpoint", "verification": "Download verified", "capabilityNote": "Adds audio clips and spoken questions to the text, image and video decision workflow. It returns structured values and probabilities, not speech. Extended answer formats require Transformers; llama.cpp supports only yes/no, choice and score. Finding boxes requires exactly one picture.", "checkpointRevision": "4cccbfd5ba3405ee0f973134bd03e0d9d60ef1d8", "apiDocumentationUrls": [ "https://huggingface.co/LiquidAI/d1-omni-2B/blob/main/LICENSE", "https://huggingface.co/blog/LiquidAI/lfm3" ], "accessNote": "Downloadable under the LFM Open License v1.0, which limits commercial use by revenue. Review the full license for your organization. Use the model\u2019s matching software and preprocessing; quality and device-speed claims on its card are publisher tests. Test confidence thresholds on examples from your own work before allowing automatic actions.", "pricingNote": "Running this download needs hardware, memory and maintenance. A model-specific hosted price and measured local operating cost are not established here.", "contextNote": "The card describes 128K context; the configuration sets 128,000 positions. The shared information, media and questions must fit that capacity; this model does not generate a reply. Questions are processed in groups of 16 by default; this is not a maximum number of questions per call.", "specialized": "decision", "questionLimit": null, "outputType": "Probabilities, choices, scores, extracted values and image boxes; no generated reply.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "liquid-d1-2b", "name": "Liquid d1 2B", "provider": "Liquid AI", "providerId": "LiquidAI/d1-2B", "sourceUrl": "https://huggingface.co/LiquidAI/d1-2B", "officialUrl": "https://huggingface.co/LiquidAI/d1-2B", "discovered": "2026-10-28", "released": "2026-10-27", "context": 128000, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": false, "note": "Route a support ticket, read a value from a document or locate objects in an image using a model you run yourself.", "deployment": "checkpoint", "verification": "Download verified", "capabilityNote": "Returns yes/no probabilities, choices, scores, extracted text, numbers, dates and image boxes. It can select more than one answer and estimate a number with a range. Extended answer formats require Transformers; llama.cpp supports only yes/no, choice and score. Finding boxes requires exactly one picture.", "checkpointRevision": "97c57047d5852d77e02dfac211f8768aff77a8b2", "apiDocumentationUrls": [ "https://huggingface.co/LiquidAI/d1-2B/blob/main/LICENSE", "https://huggingface.co/blog/LiquidAI/lfm3" ], "accessNote": "Downloadable under the LFM Open License v1.0, which limits commercial use by revenue. Review the full license for your organization. Use the model\u2019s matching software and preprocessing; quality and device-speed claims on its card are publisher tests. Test confidence thresholds on examples from your own work before allowing automatic actions.", "pricingNote": "Running this download needs hardware, memory and maintenance. A model-specific hosted price and measured local operating cost are not established here.", "contextNote": "The card describes 128K context; the configuration sets 128,000 positions. The shared information, media and questions must fit that capacity; this model does not generate a reply. Questions are processed in groups of 16 by default; this is not a maximum number of questions per call. The publisher warns that answers become less reliable beyond about 64K tokens of shared information.", "specialized": "decision", "questionLimit": null, "outputType": "Probabilities, choices, scores, extracted values and image boxes; no generated reply.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lfm3-omni-2b", "name": "LFM3 Omni 2B", "provider": "Liquid AI", "providerId": "LiquidAI/LFM3-Omni-2B", "sourceUrl": "https://huggingface.co/LiquidAI/LFM3-Omni-2B", "officialUrl": "https://huggingface.co/LiquidAI/LFM3-Omni-2B", "discovered": "2026-10-28", "released": "2026-10-27", "context": null, "maxOutput": null, "modalities": [ "text", "image", "video", "audio" ], "tools": true, "note": "Ask a spoken question about a photo, or let a local assistant see and hear a video together.", "deployment": "checkpoint", "verification": "Download verified", "capabilityNote": "Accepts text, pictures, video and audio; returns text or speech and can request tools. Your application carries out those tool calls.", "checkpointRevision": "ce2458cbce4ad970f92032ed94e3d18d30b04f9c", "apiDocumentationUrls": [ "https://huggingface.co/LiquidAI/LFM3-Omni-2B/blob/main/LICENSE", "https://huggingface.co/blog/LiquidAI/lfm3" ], "accessNote": "Downloadable under the LFM Open License v1.0, which limits commercial use by revenue. Review the full license for your organization. Use the model\u2019s matching software and preprocessing; quality and device-speed claims on its card are publisher tests.", "pricingNote": "Running this download needs hardware, memory and maintenance. A model-specific hosted price and measured local operating cost are not established here.", "contextNote": "The card specifies 32K tokens across input and output. Other internal configuration lengths do not establish a larger supported conversation; a separate output maximum is not verified.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lfm3-audio-2b", "name": "LFM3 Audio 2B", "provider": "Liquid AI", "providerId": "LiquidAI/LFM3-Audio-2B", "sourceUrl": "https://huggingface.co/LiquidAI/LFM3-Audio-2B", "officialUrl": "https://huggingface.co/LiquidAI/LFM3-Audio-2B", "discovered": "2026-10-28", "released": "2026-10-27", "context": null, "maxOutput": null, "modalities": [ "text", "audio" ], "tools": true, "note": "Build a local voice assistant that transcribes, translates, answers aloud or turns spoken requests into tool calls.", "deployment": "checkpoint", "verification": "Download verified", "capabilityNote": "Accepts text and audio and returns text, speech or both. The card lists 18 speech languages. It is not recommended for complex coding or multi-step reasoning. The card now specifies English-only audio question answering; its separate multilingual speech tasks do not imply multilingual question answering.", "checkpointRevision": "9d49701c372d6011d900f6abb40c6b296fb706dd", "apiDocumentationUrls": [ "https://huggingface.co/LiquidAI/LFM3-Audio-2B/blob/main/LICENSE", "https://huggingface.co/blog/LiquidAI/lfm3" ], "accessNote": "Downloadable under the LFM Open License v1.0, which limits commercial use by revenue. Review the full license for your organization. Use the model\u2019s matching software and preprocessing; quality and device-speed claims on its card are publisher tests.", "pricingNote": "Running this download needs hardware, memory and maintenance. A model-specific hosted price and measured local operating cost are not established here.", "contextNote": "The card specifies 32K tokens across input and output. Other internal configuration lengths do not establish a larger supported conversation; a separate output maximum is not verified.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lfm3-vl-2b", "name": "LFM3 VL 2B", "provider": "Liquid AI", "providerId": "LiquidAI/LFM3-VL-2B", "sourceUrl": "https://huggingface.co/LiquidAI/LFM3-VL-2B", "officialUrl": "https://huggingface.co/LiquidAI/LFM3-VL-2B", "discovered": "2026-10-28", "released": "2026-10-27", "context": 128000, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": true, "note": "Read documents, charts and app screens, or ask questions about a video on your own device.", "deployment": "checkpoint", "verification": "Download verified", "capabilityNote": "Supports tool calls and step-by-step thinking. The publisher does not recommend it for coding, long conversations or complex reasoning.", "checkpointRevision": "f83d2b0a2296aaa277f394990a9b8b949ddf9691", "apiDocumentationUrls": [ "https://huggingface.co/LiquidAI/LFM3-VL-2B/blob/main/LICENSE", "https://huggingface.co/blog/LiquidAI/lfm3" ], "accessNote": "Downloadable under the LFM Open License v1.0, which limits commercial use by revenue. Review the full license for your organization. Use the model\u2019s matching software and preprocessing; quality and device-speed claims on its card are publisher tests. Local GGUF use requires llama.cpp b11960 or newer and the matching vision setup.", "pricingNote": "Running this download needs hardware, memory and maintenance. A model-specific hosted price and measured local operating cost are not established here.", "contextNote": "The card describes 128K context; the configuration sets 128,000 positions. Input, media and any generated answer share the model\u2019s capacity.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lfm3-1-7b", "name": "LFM3 1.7B", "provider": "Liquid AI", "providerId": "LiquidAI/LFM3-1.7B", "sourceUrl": "https://huggingface.co/LiquidAI/LFM3-1.7B", "officialUrl": "https://huggingface.co/LiquidAI/LFM3-1.7B", "discovered": "2026-10-28", "released": "2026-10-27", "context": 262144, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Run a small text assistant on your own hardware, connect it to tools or adapt it to your business vocabulary.", "deployment": "checkpoint", "verification": "Download verified", "capabilityNote": "The publisher recommends it for local assistants and structured answers, but not complex coding agents or difficult multi-step reasoning.", "checkpointRevision": "40ebe0aa6662c9d40389c2f5d565b6a2c6aebacd", "apiDocumentationUrls": [ "https://huggingface.co/LiquidAI/LFM3-1.7B/blob/main/LICENSE", "https://huggingface.co/blog/LiquidAI/lfm3" ], "accessNote": "Downloadable under the LFM Open License v1.0, which limits commercial use by revenue. Review the full license for your organization. Use the model\u2019s matching software and preprocessing; quality and device-speed claims on its card are publisher tests.", "pricingNote": "Local costs depend on your hardware. Liquid now separately lists hosted input $0.08, cached input $0.02 and output $0.20 per million tokens; those prices do not measure the cost of running this download.", "contextNote": "The text model supports 262,144 tokens across input and output.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "rynnunify-base", "name": "RynnUnify Base \u00b7 visual planning research", "provider": "Alibaba DAMO Academy", "providerId": "Alibaba-DAMO-Academy/RynnUnify-Base", "sourceUrl": "https://huggingface.co/Alibaba-DAMO-Academy/RynnUnify-Base", "officialUrl": "https://huggingface.co/Alibaba-DAMO-Academy/RynnUnify-Base", "discovered": "2026-10-27", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": false, "note": "Explore plans for a changing physical scene by combining written reasoning with generated video predictions. This is a research setup for scene understanding and visual planning.", "deployment": "checkpoint", "verification": "Download documented", "capabilityNote": "The 16B architecture pairs an 8B reasoner with an 8B video generator. The project includes training and evaluation paths that choose whether to answer directly or generate visual evidence. Its separate world-action extension is not proof that this base download can directly control a robot.", "accessNote": "Requires the compatible Cosmos framework and project-specific setup. The card labels the checkpoint Apache 2.0; check the base-model and component terms. No ready-to-use hosted API is established.", "contextNote": "Context and maximum output limits are not verified. Video duration and research rollout examples are not general token limits.", "pricingNote": "No hosted token price or measured local cost is established. Project benchmark and robot demonstrations remain publisher evidence, not independent results here.", "apiDocumentationUrls": [ "https://github.com/alibaba-damo-academy/RynnUnify" ], "checkpointRevision": "3cea0bb1743b40b575f577443fa75c5a4487e56b", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "wevisone-vl", "name": "WeVisOne-VL \u00b7 2B and 4B", "provider": "Tencent", "providerId": "tencent/WeVisOne-VL-2B / tencent/WeVisOne-VL-4B", "sourceUrl": "https://huggingface.co/tencent/WeVisOne-VL-2B", "officialUrl": "https://huggingface.co/tencent/WeVisOne-VL-2B", "discovered": "2026-10-27", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Find, count or outline objects in a still image, locate text, or separate a subject from its background. Choose the 2B or 4B download and use Tencent\u2019s inference package.", "deployment": "checkpoint", "verification": "Download documented", "capabilityNote": "Returns structured locations and image masks. The mask decoder is a required component. This release targets one image-perception turn at a time, without an explicit thinking trace; its pixel decoder rejects video grids.", "accessNote": "Intended for visual-perception research, not identifying individuals or autonomous high-impact decisions. Boundaries can be unreliable with transparency, occlusion or low-resolution images. The repository labels Apache 2.0; third-party components retain their own terms.", "contextNote": "A usable request or maximum output allowance is not established by the card. Do not substitute the parent Qwen3-VL model\u2019s limits.", "pricingNote": "You supply compatible hardware and software. No hosted token rate or independent comparison is established.", "apiDocumentationUrls": [ "https://huggingface.co/tencent/WeVisOne-VL-4B", "https://github.com/Tencent/WeVisOne-VL" ], "checkpointRevision": "798ffc7c87ad6104410812b589e240ba32d92315", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "d3-ollama", "name": "d3 \u00b7 local decision models", "provider": "vLLM Semantic Router", "providerId": "d3:2b / d3:4b / d3:9b / d3:27b", "sourceUrl": "https://ollama.com/library/d3", "officialUrl": "https://ollama.com/library/d3", "discovered": "2026-10-27", "released": null, "context": 4096, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Route support requests, prioritize cases or check an image against a rule on your own machine. Supply the possible answers; d3 scores them instead of writing a reply.", "deployment": "checkpoint", "verification": "Download documented", "identityLabel": "Ollama package", "capabilityNote": "The four packages contain 2B, 4B, 9B and 27B variants. The plain d3 name selects 4B. Ollama supports text, structured data and still images; video and Box/Track decisions are unavailable in these packages.", "contextNote": "The package instructions specify 4,096 context tokens despite the larger 256K catalog badge. Do not assume the badge is the usable default. Publisher checkpoint capabilities and scores do not establish equivalent behavior in these compressed packages.", "accessNote": "Requires Ollama 0.40.9 or later. The page recommends Apple Silicon systems with 8GB, 16GB, 32GB and 64GB respectively, based on its default packages at 4,096 context. Other systems use a different format and may need different memory. Apache 2.0.", "pricingNote": "No per-token local API bill is established; you supply hardware and electricity. Check probabilities on real examples before setting thresholds for automatic actions.", "apiDocumentationUrls": [ "https://huggingface.co/vllm-sr/d3-nano", "https://huggingface.co/vllm-sr/d3-mini", "https://huggingface.co/vllm-sr/d3-flash", "https://huggingface.co/vllm-sr/d3" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "ocria-1", "name": "OCRia-1 \u00b7 document parsing", "provider": "LlamaIndex", "providerId": null, "sourceUrl": "https://www.llamaindex.ai/blog/parse-ocria-1", "officialUrl": "https://www.llamaindex.ai/blog/parse-ocria-1", "discovered": "2026-10-27", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Read complex document tables, charts and formatting through LlamaParse. For example, recover a financial table while keeping each value attached to its row and column.", "identityLabel": "Product model", "verification": "Provider announcement verified", "capabilityNote": "A specialized 4B model that returns Markdown and HTML, with mathematical formulas in LaTeX. It can turn charts into tables of values.", "accessNote": "Available inside LlamaParse. A standalone API model ID and public weight download are not established by the announcement.", "pricingNote": "The announcement says LlamaParse pricing is unchanged but does not give a complete model-specific tariff. Check the chosen service plan; no token price is inferred.", "dateNote": "Discovered October 27. The opened announcement did not establish an exact release date. Its benchmark comparisons were run by the provider, including on third-party datasets.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "nemotron-3-5-parse", "name": "Nemotron 3.5 Parse", "provider": "NVIDIA", "providerId": "nvidia/Nemotron-3.5-Parse", "sourceUrl": "https://huggingface.co/nvidia/Nemotron-3.5-Parse", "officialUrl": "https://huggingface.co/nvidia/Nemotron-3.5-Parse", "discovered": "2026-10-27", "released": "2026-10-17", "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Turn document images into readable text, tables and chart data while retaining reading order and locations on the page.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "62cb1e4ba153d113ff64aaf2be6e068c377cf53c", "capabilityNote": "A specialized 631-million-parameter document parser. Its helpers return Markdown, structured blocks and raw output. Extra passes revisit tables and charts; test whether they recover useful detail in your documents.", "contextNote": "The configuration has a 16,384-token sequence limit. Image processing and generated text share the model\u2019s setup; this is not a general chat input allowance.", "accessNote": "Downloadable under OpenMDW 1.1 for commercial or non-commercial use. The reference software uses Linux with Transformers or vLLM; NVIDIA reports tests on H100 hardware.", "pricingNote": "No hosted per-page or token price is established. The publisher\u2019s accuracy and pages-per-second figures use its own testing setup and are not independent results here.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "nanbeige4-3-33b", "name": "Nanbeige 4.3 33B", "provider": "Nanbeige", "providerId": "Nanbeige/Nanbeige4.3-33B", "sourceUrl": "https://huggingface.co/Nanbeige/Nanbeige4.3-33B", "officialUrl": "https://huggingface.co/Nanbeige/Nanbeige4.3-33B", "discovered": "2026-10-27", "released": null, "context": 524288, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Run a downloadable model for coding, document work and tool-assisted tasks. It is intended both for applications and for training smaller models.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "c1093c501cc6460427c0ac634db7223d0450c0ee", "capabilityNote": "Thinking is on by default and can be turned off. Keep previous thinking for multi-step tool work as the publisher recommends; ordinary chat can discard it.", "contextNote": "The publisher supports 524,288 context tokens. Its suggested answer budgets vary by task and are not a verified universal output limit.", "accessNote": "The card labels the weights Apache 2.0 and documents Transformers, SGLang and vLLM. A conventional LICENSE file was unavailable when checked. Review the repository terms before deployment.", "pricingNote": "No hosted token price or measured local running cost is established. Publisher comparisons are not independent scores here.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gpt-6-1-terra-fast-gateway", "name": "GPT-6.1 Terra Fast \u00b7 Gateway", "provider": "OpenAI", "providerId": "openai/gpt-6.1-terra-fast", "sourceUrl": "https://vercel.com/ai-gateway/models/gpt-6.1-terra-fast", "officialUrl": "https://vercel.com/ai-gateway/models/gpt-6.1-terra-fast", "discovered": "2026-10-27", "released": null, "context": 1050000, "maxOutput": 128000, "modalities": [ "text", "image" ], "tools": true, "note": "Run Terra through Gateway\u2019s faster-processing offer. Check whether reduced waiting is worth the higher bill on your own workload.", "identityLabel": "Gateway model ID", "accessNote": "A hosting option for Terra, not a new model. No independently measured speed advantage is established here.", "pricingNote": "Gateway lists $2 input, $0.10 cached input and $10 output per million tokens. Above 272,000 input tokens, input/cache rates double and output costs $15. Regional processing adds 10%. Check cache-write and tool charges separately.", "documentedRates": { "input": 2, "output": 10, "cached": 0.1, "label": "Vercel AI Gateway \u00b7 USD" }, "supportedEfforts": [ "none", "low", "medium", "high", "xhigh", "max" ], "effortScope": "Vercel AI Gateway", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gyro-qwen3-8-flash-next", "name": "Gyro S / M / L \u00b7 local variants", "provider": "AgentionAI", "providerId": "agentionai/Qwen3.8-Flash-Next-Gyro-GGUF", "sourceUrl": "https://huggingface.co/agentionai/Qwen3.8-Flash-Next-Gyro-GGUF", "officialUrl": "https://huggingface.co/agentionai/Qwen3.8-Flash-Next-Gyro-GGUF", "discovered": "2026-10-26", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Choose a smaller download format for running Qwen3.8 Flash Next locally. S, M and L offer different memory requirements; compare them on your own mix of writing and coding.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "05892758a67efff0895d23c6df10a6fdb6be8cf4", "capabilityNote": "S, M and L are compressed versions of the same base model, not three new base releases. Image input needs the separate vision file and a compatible runtime. Older and newer files can share a name: save the repository revision with your tests.", "contextNote": "Publisher GPU-weight figures are 27.64 GiB for S, 34.98 GiB for M and 50.21 GiB for L. Allow additional memory for the request, vision components and optional draft model; the weight size alone does not tell you what fits.", "accessNote": "Use a compatible AgentionAI, Strata or strix-llama.cpp build. Standard llama.cpp cannot load the format, and support differs by variant. The Qwen Community License includes extra conditions for commercial model hosting and coding/work-assistant businesses.", "pricingNote": "Local hardware and electricity costs depend on your setup. The large lookup table can stay on a fast NVMe drive, so download size is greater than the GPU-weight figure. Vendor speed and quality claims are not independent scores here.", "apiDocumentationUrls": [ "https://huggingface.co/agentionai/Qwen3.8-Flash-Next-Gyro-GGUF/blob/main/LICENSE" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gyro-coder", "name": "Gyro-Coder \u00b7 local coding variants", "provider": "AgentionAI", "providerId": "agentionai/Qwen3.8-Flash-Next-Gyro-Coder-GGUF", "sourceUrl": "https://huggingface.co/agentionai/Qwen3.8-Flash-Next-Gyro-Coder-GGUF", "officialUrl": "https://huggingface.co/agentionai/Qwen3.8-Flash-Next-Gyro-Coder-GGUF", "discovered": "2026-10-26", "released": "2026-10-21", "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Try local coding on a 24GB or 32GB graphics card. These compressed, coding-focused versions of Qwen3.8 Flash Next trade broader knowledge for a smaller memory requirement.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "b7d23f5036e9134b3b84e20e7040ba47bc5e0b4c", "capabilityNote": "Choose Coder for a 24GB card or Coder-L for 32GB. Give clear tasks such as fixing a specific bug or implementing a function. They retain half of the original model\u2019s specialist components; they are variants, not new base models.", "contextNote": "The publisher\u2019s Coder example uses 65,536 context tokens and about 21.1 GiB of GPU memory. Longer requests and other settings can need more memory; this example is not a maximum context guarantee.", "accessNote": "Requires AgentionAI\u2019s compatible runtime; standard llama.cpp cannot load these files. The Qwen Community License has additional conditions for commercial model hosting and coding/work-assistant businesses. Check those terms before offering a service. Coder-L v1.2 replaces the file with the same name on October 29 to address tool-call formatting and symbols. Save the repository revision; the reported improvements are publisher tests, not a new base-model release.", "pricingNote": "Coder takes 70.6 GiB on disk despite using 19.92 GiB for GPU weights; Coder-L takes 78.2 GiB on disk and 27.54 GiB for GPU weights. Keep the large lookup table on a fast solid-state drive (NVMe). These are publisher figures, not our hardware tests.", "apiDocumentationUrls": [ "https://huggingface.co/agentionai/Qwen3.8-Flash-Next-Gyro-GGUF", "https://huggingface.co/agentionai/Qwen3.8-Flash-Next-Gyro-Coder-GGUF/blob/main/LICENSE" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "flora-captioner", "name": "Flora Captioner \u00b7 video labels", "provider": "NVIDIA", "providerId": "nvidia/Flora-Captioner", "sourceUrl": "https://huggingface.co/nvidia/Flora-Captioner", "officialUrl": "https://huggingface.co/nvidia/Flora-Captioner", "discovered": "2026-10-26", "released": "2026-10-15", "context": null, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": false, "note": "Describe an action in a selected part of a robot or human demonstration video. For example, turn a clip into an instruction that can label a training example.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "ad744f46335aec671f5caac09ac97d817fbdb114", "capabilityNote": "You choose the start and end of the action. The model uses that clip and nearby frames to write one English instruction; it does not find or change the boundaries. The download contains the full 27B model.", "contextNote": "Follow the publisher\u2019s frame-sampling and timing instructions. Its reference setup uses a 32,768-token sequence budget; this is not a verified general-purpose context limit.", "accessNote": "For non-commercial robotics research and development only, under the NVIDIA OneWay Noncommercial License. These tools label videos; they are not supported for production use, direct robot control or use as the sole safety monitor.", "pricingNote": "Downloadable research model; no hosted token price is established.", "apiDocumentationUrls": [ "https://developer.download.nvidia.com/licenses/NVIDIA-OneWay-Noncommercial-License-22Mar2022.pdf" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "flora-tempseg", "name": "Flora TempSeg \u00b7 video boundaries", "provider": "NVIDIA", "providerId": "nvidia/Flora-TempSeg", "sourceUrl": "https://huggingface.co/nvidia/Flora-TempSeg", "officialUrl": "https://huggingface.co/nvidia/Flora-TempSeg", "discovered": "2026-10-26", "released": "2026-10-15", "context": null, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": false, "note": "Find where one action ends and another begins in a robot or human demonstration video. Use the results to prepare a robotics research dataset.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "b3f079b160a23f78026b72b80fb7a0e05c57eba8", "capabilityNote": "Returns time intervals and flags for actions that continue beyond a clip; it does not name the actions. This download is an adapter: it also needs the exact Qwen3.5-9B base specified in the model card. The reference setup processes 16- or 32-second windows.", "contextNote": "Follow the publisher\u2019s frame-sampling and timing instructions. Its reference setup uses a 32,768-token sequence budget; this is not a verified general-purpose context limit.", "accessNote": "For non-commercial robotics research and development only, under the NVIDIA OneWay Noncommercial License. These tools label videos; they are not supported for production use, direct robot control or use as the sole safety monitor.", "pricingNote": "Downloadable research model; no hosted token price is established.", "apiDocumentationUrls": [ "https://developer.download.nvidia.com/licenses/NVIDIA-OneWay-Noncommercial-License-22Mar2022.pdf" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "openmobile-2-27b", "name": "OpenMobile 2 27B", "provider": "OpenMobile-2", "providerId": "OpenMobile-2/OpenMobile-2-27B", "sourceUrl": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B", "officialUrl": "https://huggingface.co/OpenMobile-2/OpenMobile-2-27B", "discovered": "2026-10-26", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": true, "note": "Build an Android assistant that reads screenshots and takes actions in apps. Your application connects the model to the phone and decides which actions it may execute.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "3d844db8c20421524a5012837b151e444ad35ae6", "capabilityNote": "Can select screen actions and call app-specific tools when the app currently open makes them available. Use the matching agent software and test complete workflows, including mistakes and recovery.", "contextNote": "The publisher\u2019s serving examples use a 65,536-token input-and-answer budget. That example does not establish the model\u2019s maximum supported context or output.", "accessNote": "Downloadable weights are labelled Apache 2.0. The 9B version builds on Qwen3.5-9B; 27B builds on Qwen3.6-27B. Their scores cannot be substituted for one another or for the parent models.", "pricingNote": "No hosted token price is established. For local use, include hardware, electricity and setup time in your budget.", "apiDocumentationUrls": [ "https://github.com/OpenMobile2/OpenMobile2-Code", "https://os-copilot.github.io/OpenMobile2-Home/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "openmobile-2-9b", "name": "OpenMobile 2 9B", "provider": "OpenMobile-2", "providerId": "OpenMobile-2/OpenMobile-2-9B", "sourceUrl": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B", "officialUrl": "https://huggingface.co/OpenMobile-2/OpenMobile-2-9B", "discovered": "2026-10-26", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": true, "note": "Build an Android assistant that reads screenshots and takes actions in apps. Your application connects the model to the phone and decides which actions it may execute.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "68ebbac4d5bdf725103a22d076122b8aa883a0f2", "capabilityNote": "Can select screen actions and call app-specific tools when the app currently open makes them available. Use the matching agent software and test complete workflows, including mistakes and recovery.", "contextNote": "The publisher\u2019s serving examples use a 65,536-token input-and-answer budget. That example does not establish the model\u2019s maximum supported context or output.", "accessNote": "Downloadable weights are labelled Apache 2.0. The 9B version builds on Qwen3.5-9B; 27B builds on Qwen3.6-27B. Their scores cannot be substituted for one another or for the parent models.", "pricingNote": "No hosted token price is established. For local use, include hardware, electricity and setup time in your budget.", "apiDocumentationUrls": [ "https://github.com/OpenMobile2/OpenMobile2-Code", "https://os-copilot.github.io/OpenMobile2-Home/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jais-3-375b-a23b", "name": "Jais 3 375B-A23B", "provider": "Inception / IFM", "providerId": "inception42/Jais-3-375B-A23B", "sourceUrl": "https://huggingface.co/inception42/Jais-3-375B-A23B", "officialUrl": "https://huggingface.co/inception42/Jais-3-375B-A23B", "discovered": "2026-10-25", "released": "2026-10-15", "context": 524288, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Build an Arabic and English text assistant, including regional Arabic and Arabizi. Test it on the dialects, documents and tool workflows your users actually need.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "69a96a885e93367c508c0428d01b5c38d23bdd1d", "capabilityNote": "The provider recommends high thinking effort. Low is a research mode with reasoning disabled; the default is not confirmed. Your application executes any tools the model requests.", "contextNote": "The documented capacity is 524,288 tokens. Prompt, thinking and answer share this space; a maximum answer length is not established.", "accessNote": "The publisher labels the downloadable files Apache 2.0. Use its matching K2-Horizon software and choose a model size that fits your hardware, leaving memory for the requests too.", "pricingNote": "Budget for hardware, electricity and operation. No hosted token price or measured local running cost is confirmed.", "apiDocumentationUrls": [ "https://ifm.ai/blog/jais3/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jais-3-mova-36b-a4b", "name": "Jais 3 MoVA-36B-A4B", "provider": "Inception / IFM", "providerId": "inception42/Jais-3-MoVA-36B-A4B", "sourceUrl": "https://huggingface.co/inception42/Jais-3-MoVA-36B-A4B", "officialUrl": "https://huggingface.co/inception42/Jais-3-MoVA-36B-A4B", "discovered": "2026-10-25", "released": "2026-10-15", "context": 524288, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Build an Arabic and English text assistant, including regional Arabic and Arabizi. Test it on the dialects, documents and tool workflows your users actually need.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "83e05890681998bafd06b082e8f8ddd1241e0de2", "capabilityNote": "The provider recommends high thinking effort. Low is a research mode with reasoning disabled; the default is not confirmed. Your application executes any tools the model requests.", "contextNote": "The documented capacity is 524,288 tokens. Prompt, thinking and answer share this space; a maximum answer length is not established.", "accessNote": "The publisher labels the downloadable files Apache 2.0. Use its matching K2-Horizon software and choose a model size that fits your hardware, leaving memory for the requests too.", "pricingNote": "Budget for hardware, electricity and operation. No hosted token price or measured local running cost is confirmed.", "apiDocumentationUrls": [ "https://ifm.ai/blog/jais3/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jais-3-32b", "name": "Jais 3 32B", "provider": "Inception / IFM", "providerId": "inception42/Jais-3-32B", "sourceUrl": "https://huggingface.co/inception42/Jais-3-32B", "officialUrl": "https://huggingface.co/inception42/Jais-3-32B", "discovered": "2026-10-25", "released": "2026-10-15", "context": 524288, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Build an Arabic and English text assistant, including regional Arabic and Arabizi. Test it on the dialects, documents and tool workflows your users actually need.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "170144803abac94000f966ec215596d78a9bbcd8", "capabilityNote": "The provider recommends high thinking effort. Low is a research mode with reasoning disabled; the default is not confirmed. Your application executes any tools the model requests.", "contextNote": "The documented capacity is 524,288 tokens. Prompt, thinking and answer share this space; a maximum answer length is not established.", "accessNote": "The publisher labels the downloadable files Apache 2.0. Use its matching K2-Horizon software and choose a model size that fits your hardware, leaving memory for the requests too.", "pricingNote": "Budget for hardware, electricity and operation. No hosted token price or measured local running cost is confirmed.", "apiDocumentationUrls": [ "https://ifm.ai/blog/jais3/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jais-3-7b", "name": "Jais 3 7B", "provider": "Inception / IFM", "providerId": "inception42/Jais-3-7B", "sourceUrl": "https://huggingface.co/inception42/Jais-3-7B", "officialUrl": "https://huggingface.co/inception42/Jais-3-7B", "discovered": "2026-10-25", "released": "2026-10-15", "context": 524288, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Build an Arabic and English text assistant, including regional Arabic and Arabizi. Test it on the dialects, documents and tool workflows your users actually need.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "5a1f6ce4cc6427f6988d5631ab5b204b78bba023", "capabilityNote": "The provider recommends high thinking effort. Low is a research mode with reasoning disabled; the default is not confirmed. Your application executes any tools the model requests.", "contextNote": "The documented capacity is 524,288 tokens. Prompt, thinking and answer share this space; a maximum answer length is not established.", "accessNote": "The publisher labels the downloadable files Apache 2.0. Use its matching K2-Horizon software and choose a model size that fits your hardware, leaving memory for the requests too.", "pricingNote": "Budget for hardware, electricity and operation. No hosted token price or measured local running cost is confirmed.", "apiDocumentationUrls": [ "https://ifm.ai/blog/jais3/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jais-3-3-7b", "name": "Jais 3 3.7B", "provider": "Inception / IFM", "providerId": "inception42/Jais-3-3.7B", "sourceUrl": "https://huggingface.co/inception42/Jais-3-3.7B", "officialUrl": "https://huggingface.co/inception42/Jais-3-3.7B", "discovered": "2026-10-25", "released": "2026-10-15", "context": 524288, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Build an Arabic and English text assistant, including regional Arabic and Arabizi. Test it on the dialects, documents and tool workflows your users actually need.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "351b5ad88efe27531926621931a069b4a9e0d09d", "capabilityNote": "The provider recommends high thinking effort. Low is a research mode with reasoning disabled; the default is not confirmed. Your application executes any tools the model requests.", "contextNote": "The documented capacity is 524,288 tokens. Prompt, thinking and answer share this space; a maximum answer length is not established.", "accessNote": "The publisher labels the downloadable files Apache 2.0. Use its matching K2-Horizon software and choose a model size that fits your hardware, leaving memory for the requests too.", "pricingNote": "Budget for hardware, electricity and operation. No hosted token price or measured local running cost is confirmed.", "apiDocumentationUrls": [ "https://ifm.ai/blog/jais3/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jais-3-0-9b", "name": "Jais 3 0.9B", "provider": "Inception / IFM", "providerId": "inception42/Jais-3-0.9B", "sourceUrl": "https://huggingface.co/inception42/Jais-3-0.9B", "officialUrl": "https://huggingface.co/inception42/Jais-3-0.9B", "discovered": "2026-10-25", "released": "2026-10-15", "context": 131072, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Build an Arabic and English text assistant, including regional Arabic and Arabizi. Test it on the dialects, documents and tool workflows your users actually need.", "deployment": "checkpoint", "verification": "Download verified", "checkpointRevision": "b2d9e4ea4fddd86b781dc350ed91d4ccc7acfb41", "capabilityNote": "The provider recommends high thinking effort. Low is a research mode with reasoning disabled; the default is not confirmed. Your application executes any tools the model requests.", "contextNote": "The documented capacity is 131,072 tokens. Prompt, thinking and answer share this space; a maximum answer length is not established.", "accessNote": "The publisher labels the downloadable files Apache 2.0. Use its matching K2-Horizon software and choose a model size that fits your hardware, leaving memory for the requests too.", "pricingNote": "Budget for hardware, electricity and operation. No hosted token price or measured local running cost is confirmed.", "apiDocumentationUrls": [ "https://ifm.ai/blog/jais3/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "kettei-dot", "name": "kettei dot \u00b7 announced", "provider": "Sqwish", "providerId": null, "sourceUrl": "https://sqwish.ai/blog/your-ai-agent-is-writing-essays-to-tick-boxes", "officialUrl": "https://sqwish.ai/blog/your-ai-agent-is-writing-essays-to-tick-boxes", "discovered": "2026-10-25", "released": null, "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "An announced small text decision model intended for local use. A usable download and public API ID are not yet verified.", "announced": "2026-10-14", "availability": "announced", "verification": "Announced; download pending verification", "specialized": "decision", "outputType": "Fixed choices and scores", "accessNote": "The publisher describes a small text model based on K2-Horizon and promises open-source access. Its current API offers fat, chub and core only; a usable dot download is not yet confirmed.", "pricingNote": "No current hosted price or usable download is confirmed. Wait for access details before planning to use this model.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "kettei-core", "name": "kettei core \u00b7 Sqwish API", "provider": "Sqwish", "providerId": "kettei-core", "sourceUrl": "https://docs.sqwish.ai/guides/models-and-fallback", "officialUrl": "https://docs.sqwish.ai/guides/models-and-fallback", "discovered": "2026-10-25", "released": "2026-10-14", "context": 65536, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Route a request, check a record or choose an action from answers you define. Use your own labeled examples to choose the smallest model that works well enough.", "verification": "Native API verified", "specialized": "decision", "outputType": "Yes/no probabilities, choices and scores; no generated explanation.", "questionLimit": 128, "documentedRates": { "input": 0.02, "output": 0, "cached": null, "label": "Sqwish base model \u00b7 USD" }, "contextNote": "Up to 65,536 billed input tokens and 128 questions per base-model request. Fine-tuned models have a separate 32-question limit.", "capabilityNote": "Accepts text; this model does not read images. Automatic fallback is on by default and can choose another base model. Use fallback: none for an exact-model test, and record the returned model and version.", "pricingNote": "Base input costs $0.02 per million tokens; output is free. Fine-tuned rates differ. A fallback is billed at the serving rate, capped at the requested model\u2019s rate.", "accessNote": "Check account access and model readiness before relying on it. Publisher benchmark and speed claims are not independent results here; downloadable core weights are not yet verified. Optional zero data retention for inference requires an organization agreement and is off by default. It excludes stored usage records and training data.", "apiDocumentationUrls": [ "https://docs.sqwish.ai/pricing", "https://api.sqwish.ai/openapi.json", "https://sqwish.ai/blog/your-ai-agent-is-writing-essays-to-tick-boxes", "https://docs.sqwish.ai/guides/zero-data-retention" ], "billingNote": " With zero data retention enabled, repeated requests run and are billed again, even with the same retry key. Include automatic SDK retries in your cost estimate.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "kettei-chub", "name": "kettei chub \u00b7 Sqwish API", "provider": "Sqwish", "providerId": "kettei-chub", "sourceUrl": "https://docs.sqwish.ai/guides/models-and-fallback", "officialUrl": "https://docs.sqwish.ai/guides/models-and-fallback", "discovered": "2026-10-25", "released": "2026-10-14", "context": 65536, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Route a request, check a record or choose an action from answers you define. Use your own labeled examples to choose the smallest model that works well enough.", "verification": "Native API verified", "specialized": "decision", "outputType": "Yes/no probabilities, choices and scores; no generated explanation.", "questionLimit": 128, "documentedRates": { "input": 0.04, "output": 0, "cached": null, "label": "Sqwish base model \u00b7 USD" }, "contextNote": "Up to 65,536 billed input tokens and 128 questions per base-model request. Fine-tuned models have a separate 32-question limit. Image tokens count toward the same input limit.", "capabilityNote": "Accepts text and up to eight images. Automatic fallback is on by default and can choose another base model. Use fallback: none for an exact-model test, and record the returned model and version.", "pricingNote": "Base input costs $0.04 per million tokens; output is free. Fine-tuned rates differ. A fallback is billed at the serving rate, capped at the requested model\u2019s rate.", "accessNote": "Check account access and model readiness before relying on it. API access does not establish a downloadable model. Test accuracy and response time on your own examples. Optional zero data retention for inference requires an organization agreement and is off by default. It excludes stored usage records and training data.", "apiDocumentationUrls": [ "https://docs.sqwish.ai/pricing", "https://api.sqwish.ai/openapi.json", "https://sqwish.ai/blog/your-ai-agent-is-writing-essays-to-tick-boxes", "https://docs.sqwish.ai/guides/zero-data-retention", "https://docs.sqwish.ai/guides/images" ], "billingNote": "With zero data retention enabled, repeated requests run and are billed again, even with the same retry key. Include automatic SDK retries in your cost estimate. Each unique image now adds 64\u20132,048 billed input tokens according to size, not a flat 64. Use the estimate endpoint or returned input usage.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "kettei-fat", "name": "kettei fat \u00b7 Sqwish API", "provider": "Sqwish", "providerId": "kettei-fat", "sourceUrl": "https://docs.sqwish.ai/guides/models-and-fallback", "officialUrl": "https://docs.sqwish.ai/guides/models-and-fallback", "discovered": "2026-10-25", "released": "2026-10-14", "context": 65536, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Route a request, check a record or choose an action from answers you define. Use your own labeled examples to choose the smallest model that works well enough.", "verification": "Native API verified", "specialized": "decision", "outputType": "Yes/no probabilities, choices and scores; no generated explanation.", "questionLimit": 128, "documentedRates": { "input": 0.06, "output": 0, "cached": null, "label": "Sqwish base model \u00b7 USD" }, "contextNote": "Up to 65,536 billed input tokens and 128 questions per base-model request. Fine-tuned models have a separate 32-question limit. Image tokens count toward the same input limit.", "capabilityNote": "Accepts text and up to eight images. Automatic fallback is on by default and can choose another base model. Use fallback: none for an exact-model test, and record the returned model and version.", "pricingNote": "Base input costs $0.06 per million tokens; output is free. Fine-tuned rates differ. A fallback is billed at the serving rate, capped at the requested model\u2019s rate.", "accessNote": "Check account access and model readiness before relying on it. API access does not establish a downloadable model. Test accuracy and response time on your own examples. Optional zero data retention for inference requires an organization agreement and is off by default. It excludes stored usage records and training data.", "apiDocumentationUrls": [ "https://docs.sqwish.ai/pricing", "https://api.sqwish.ai/openapi.json", "https://sqwish.ai/blog/your-ai-agent-is-writing-essays-to-tick-boxes", "https://docs.sqwish.ai/guides/zero-data-retention", "https://docs.sqwish.ai/guides/images" ], "billingNote": "With zero data retention enabled, repeated requests run and are billed again, even with the same retry key. Include automatic SDK retries in your cost estimate. Each unique image now adds 64\u20132,048 billed input tokens according to size, not a flat 64. Use the estimate endpoint or returned input usage.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "liquid-d1-1", "name": "Liquid d1.1 \u00b7 native API", "provider": "Liquid AI", "providerId": "d1.1", "sourceUrl": "https://docs.liquid.ai/lfm/models/d1", "officialUrl": "https://docs.liquid.ai/lfm/models/d1", "discovered": "2026-10-25", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Classify a record or image, choose from allowed answers, or score a case against your rules. This is now Liquid\u2019s default hosted decision model.", "specialized": "decision", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "verification": "Native API verified", "documentedRates": { "input": 0.02, "output": 0, "cached": 0.02, "label": "Liquid native API \u00b7 USD" }, "capabilityNote": "Each question is billed for its complete text and all supplied images. Test confidence thresholds on examples from your own task.", "contextNote": "An exact context limit and maximum question count are not verified.", "pricingNote": "Liquid now explicitly lists d1.1 at $0.02 per million billed input tokens, with no output charge. Cached input has the same $0.02 rate, so it does not provide a token-price discount.", "accessNote": "Use the exact d1.1 ID. Earlier d1 remains separately listed on OpenRouter and Gateway; no retirement date for it is established.", "apiDocumentationUrls": [ "https://www.liquid.ai/pricing" ], "billingNote": "Each question is charged for its complete text and all supplied images. Use the API\u2019s billed input total; multiple questions can repeat input costs.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "kimi-k3-1-pro-highspeed", "name": "Kimi K3.1 Pro Highspeed \u00b7 Gateway", "provider": "Moonshot AI", "providerId": "moonshotai/kimi-k3.1-pro-highspeed", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "officialUrl": "https://ai-gateway.vercel.sh/v1/models", "discovered": "2026-10-25", "released": "2026-10-23", "context": 1048576, "maxOutput": 1048576, "modalities": [ "text", "image" ], "tools": true, "note": "Consider the higher-priced service when response time matters. Test the time saved on your real requests before paying more.", "identityLabel": "AI Gateway model ID", "verification": "Hosting offer verified", "documentedRates": { "input": 4, "cached": 0.2, "output": 20, "write": 4, "label": "AI Gateway \u00b7 USD" }, "contextNote": "Gateway declares 1,048,576 context and maximum output tokens. Prompt, thinking and answer must still fit the shared context.", "capabilityNote": "Gateway documents low, high and max reasoning effort, but does not state an omitted default. Native Kimi defaults to max and keeps thinking on. Check the selected service\u2019s settings.", "pricingNote": "Per million tokens: $4 input, $0.20 cache reads, $20 output and $4 cache writes. Standard K3.1 Pro on Gateway lists $2/$0.20/$10/$2. Highspeed doubles the input and output quote, not the cache-read quote.", "accessNote": "This is a serving option for K3.1 Pro, not a separately benchmarked base model. The publisher\u2019s speed claim is not a measured advantage for your workload.", "apiDocumentationUrls": [ "https://platform.kimi.ai/docs/guide/kimi-k3-1-pro-quickstart" ], "supportedEfforts": [ "low", "high", "max" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "nexus-auto", "name": "Nexus Auto", "provider": "Fireworks", "providerId": "auto", "sourceUrl": "https://docs.fireworks.ai/nexus/firerouter", "officialUrl": "https://docs.fireworks.ai/nexus/firerouter", "discovered": "2026-10-23", "released": null, "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Let Fireworks choose a model for each turn of coding work. If you need one fixed model, check whether your plan supports pinning it.", "verification": "Renamed routing service verified", "capabilityNote": "Replaces FireRouter. Open models bill to Fireworks; Anthropic and OpenAI turns use your connected credentials. The response model field identifies what answered. Without a closed-provider credential, that provider is left out of automatic selection.", "accessNote": "With a Fireworks key alone, auto uses open models. Connected Anthropic/OpenAI credentials add those providers. The current guide limits pinning a specific model to Enterprise; confirm your plan before relying on fixed selection.", "pricingNote": "No single token price or guaranteed saving. Compare the total bill for completed tasks, including retries and the model selected for each turn.", "contextNote": "Limits and reasoning behavior depend on the serving model; a common exact maximum is not verified.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "squad-0-8b-labqoat", "name": "squad 0.8B \u00b7 Labqoat", "provider": "Labqoat", "providerId": "labqoat/squad-0.8b", "sourceUrl": "https://huggingface.co/labqoat/squad-0.8b", "officialUrl": "https://huggingface.co/labqoat/squad-0.8b", "discovered": "2026-10-23", "released": null, "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Make local choices for games, web-app testing, multilingual tasks and documents. This revision changes the underlying model as well as adding specialists, so retest existing workflows before upgrading.", "deployment": "checkpoint", "verification": "Download verified", "specialized": "decision", "outputType": "Choices, yes/no probabilities and ordered scores; no written answer.", "questionLimit": null, "checkpointRevision": "4147aeb09b3bb2bf3335c2062d179db0b3e7fc7b", "contextNote": "A validated task input limit is not established here. The base model\u2019s configured capacity is not a guarantee for long testing workflows.", "capabilityNote": "The publisher revised its web-app specialist again and reports improved tests. These remain publisher evaluations; retest your own complete workflow before replacing an earlier download.", "accessNote": "Apache 2.0-labelled weights, with CUDA, MLX and GGUF paths. The publisher still reports poor reliability on real web apps. Image input was not re-evaluated after the base model changed; test it before relying on it.", "pricingNote": "No hosted token price. Check whether its confidence numbers match actual success on your examples; they were adjusted using the previous version. The publisher\u2019s tests do not establish independent performance, and a web-testing setting was chosen using those same tests.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lemondecider-4b", "name": "LemonDecider 4B", "provider": "InclusionAI", "providerId": "inclusionAI/LemonDecider", "sourceUrl": "https://huggingface.co/inclusionAI/LemonDecider", "officialUrl": "https://huggingface.co/inclusionAI/LemonDecider", "discovered": "2026-10-23", "released": null, "context": 8192, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Classify text locally using fixed choices or a scoring scale you define. This v3 package splits long question lists while keeping the full shared information in each part.", "deployment": "checkpoint", "verification": "Download verified", "specialized": "decision", "outputType": "Choices, yes/no probabilities and ordered scores; no generated explanation.", "questionLimit": null, "checkpointRevision": "81556cfa823e315d7ed3e027e44905c4cce45398", "contextNote": "Each complete chunk, including formatting, must fit in 8,192 tokens. Longer question lists are split recursively, repeating the full shared information and all options; inputs are not truncated. The older 262,144-token claim does not describe this v3 service.", "capabilityNote": "The current package identifies lemondecider-4b-20261023-v3. It returns choices, yes/no probabilities and ordered scores, with up to 255 options per question. The publisher tests how closely confidence scores match correct answers; check this on your own examples before using a score to trigger actions.", "accessNote": "Apache 2.0-labelled weights. Use the supplied decision runtime to load the model and return probabilities.", "pricingNote": "No hosted price verified. The publisher\u2019s own evaluation runs are not official leaderboard submissions or independent results.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "ling-3-1-flash-1019", "name": "Ling 3.1 Flash 1019", "provider": "InclusionAI", "providerId": "inclusionAI/Ling-3.1-flash-1019", "sourceUrl": "https://x.com/AntLingAGI/status/2113226531522891851", "officialUrl": "https://x.com/AntLingAGI/status/2113226531522891851", "discovered": "2026-10-23", "released": "2026-10-27", "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": null, "note": "Download this revised Ling model for coding and tool-based workflows. Test the exact revision before replacing the earlier Flash model.", "announced": "2026-10-22", "availability": "available", "verification": "Download verified", "accessNote": "The publisher announced downloadable weights October 27. The model card labels them MIT; check the license included with your chosen package before reuse.", "pricingNote": "No verified current price. Earlier Ling prices and scores are not assigned to this revision.", "capabilityNote": "The publisher recommends enabling thinking. Its performance claims are not independent results here.", "deployment": "checkpoint", "checkpointRevision": "dddfee4bb43519a92761fbc5cefe3f0d3ff4aded", "contextNote": "The card advertises a 1M-token window, while this download configuration sets 262,144 positions. Confirm the supported serving setup before relying on a million-token request.", "apiDocumentationUrls": [ "https://huggingface.co/inclusionAI/Ling-3.1-flash-1019" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "glm-5-3-flashx-legal", "name": "GLM 5.3 FlashX Legal", "provider": "Z.ai", "providerId": "glm-5.3-flashx-legal", "sourceUrl": "https://docs.z.ai/guides/vlm/glm-5.3-flash", "officialUrl": "https://docs.z.ai/guides/vlm/glm-5.3-flash", "discovered": "2026-10-23", "released": null, "context": 1000000, "maxOutput": 128000, "modalities": [ "text", "image", "video", "file" ], "tools": true, "note": "A specialized option for legal research and document workflows in English and Chinese. Check its answers against the underlying documents before relying on them.", "verification": "Provider API ID verified", "documentedRates": { "input": 0.37, "cached": 0.075, "output": 1.25, "label": "Z.ai native API \u00b7 USD" }, "contextNote": "The shared Flash/FlashX guide lists 1M context and 128K maximum output. The catalog identifies Legal as a specialized model, not a new measurement of the general Flash model.", "capabilityNote": "The guide requires thinking to stay on and recommends max effort. It does not say which setting applies if you leave it out.", "pricingNote": "Native prices per million tokens: $0.37 input, $0.075 cached input and $1.25 output. Cache storage is currently marked limited-time free.", "accessNote": "The native guide no longer limits access to selected enterprise applicants. OpenRouter and Gateway now list this exact Legal variant; use the ID and limits of your chosen service. Downloadable Legal weights and independent legal-task results remain unverified.", "apiDocumentationUrls": [ "https://docs.z.ai/guides/overview/overview", "https://docs.z.ai/guides/overview/pricing", "https://openrouter.ai/z-ai/glm-5.3-flashx-legal", "https://vercel.com/ai-gateway/models/glm-5.3-flashx-legal" ], "catalogOffer": { "id": "z-ai/glm-5.3-flashx-legal", "input": 0.37, "output": 1.25, "cached": 0.075, "checked": "2026-11-03", "observed": "2026-11-03", "sourceUrl": "https://openrouter.ai/z-ai/glm-5.3-flashx-legal", "context": 1048576, "maxOutput": 131072, "note": "OpenRouter declares 1,048,576 context and 131,072 output, with mandatory thinking and max default. Gateway separately declares 1,000,000/131,072 and low/high/max controls without a default. These do not replace native 1M/128K documentation." }, "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "mammoth-1-local", "name": "Mammoth-1 \u00b7 download", "provider": "Baseten / Base Labs", "providerId": "baselabs/mammoth-1", "sourceUrl": "https://huggingface.co/baselabs/mammoth-1", "officialUrl": "https://huggingface.co/baselabs/mammoth-1", "discovered": "2026-10-23", "released": "2026-10-21", "context": 16384, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Run fixed-choice classification and scoring on your own hardware, with text or images.", "deployment": "checkpoint", "verification": "Download verified", "specialized": "decision", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "questionLimit": null, "contextNote": "The model card documents 16,384 maximum sequence length. Check the exact service if you use a hosted copy; a local limit does not guarantee that service\u2019s limits.", "accessNote": "MIT-labelled 27B weights and a pinned CUDA serving setup. The vision encoder is frozen. Check memory for all weights plus request data, and validate confidence on your own examples.", "pricingNote": "The hosted launch promotion does not pay for your local hardware. No measured local operating cost or independent comparison is established.", "checkpointRevision": "61cec44c281c693992e3d6ebeb44dd3402dfa186", "capabilityNote": "Supply text or images and the choices you want the model to compare. Your application must carry out any action you choose from its answer.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "mammoth-1-gateway", "name": "Mammoth-1 \u00b7 Gateway", "provider": "Baseten / Base Labs", "providerId": "baseten/mammoth-1", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "officialUrl": "https://ai-gateway.vercel.sh/v1/models", "discovered": "2026-10-23", "released": "2026-10-21", "context": 16384, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Choose from defined answers, route records or score a condition. Compare this model on your own examples before automating decisions.", "specialized": "decision", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "questionLimit": null, "identityLabel": "AI Gateway model", "verification": "Hosting offer verified", "documentedRates": { "input": 0.1, "output": 0, "cached": null, "label": "AI Gateway \u00b7 USD" }, "contextNote": "Gateway declares 16,384 tokens and text input. This service returns decisions rather than a generated written answer.", "capabilityNote": "The catalog documents choices, yes/no probabilities and ordered scores. Native image support and native question limits must not be assumed for this hosted offer.", "pricingNote": "The launch promotion has ended. Current standard input is $0.1 per million billed tokens, with no output-token charge. Native service-tier discounts do not automatically apply to Gateway.", "billingNote": "Use actual billed input tokens, including the questions. The current paid rate replaces the launch promotion.", "apiDocumentationUrls": [ "https://vercel.com/changelog/basetens-mammoth-1-and-elephant-1-now-available-on-ai-gateway-for-free" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "mammoth-1", "name": "Mammoth-1 \u00b7 Baseten API", "provider": "Baseten / Base Labs", "providerId": "decision-models/baseten-mammoth-27b-latest", "sourceUrl": "https://docs.baseten.co/inference/model-apis/decision-models", "officialUrl": "https://docs.baseten.co/inference/model-apis/decision-models", "discovered": "2026-10-23", "released": "2026-10-21", "context": 16384, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Choose from defined answers, route records or score a condition. Compare this model on your own examples before automating decisions.", "specialized": "decision", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "questionLimit": null, "verification": "Native API verified", "documentedRates": { "input": 0.1, "output": 0, "cached": null, "label": "Native Baseten API \u00b7 USD" }, "contextNote": "Each question\u2019s full prompt must fit 16,384 tokens, including shared input, options, formatting and images. Over-limit requests are rejected rather than shortened.", "capabilityNote": "Up to four images as base64 data URLs or supported cloud-storage HTTPS URLs. Questions are evaluated independently. Shared input is counted once per request, plus each question; confidence measures how concentrated the answer probabilities are.", "pricingNote": "The launch promotion has ended. Current standard input is $0.1 per million billed tokens, with no output-token charge. Native Flex is approximately 90% of default; priority defaults to twice the standard rate.", "billingNote": "Use actual billed input tokens, including the questions. The current paid rate replaces the launch promotion.", "accessNote": "Use the exact Baseten ID. Gateway\u2019s limits and modality tags differ. The publisher\u2019s evaluation claims are not independent scores here. Flex and priority are optional processing settings; faster completion is not independently measured here.", "apiDocumentationUrls": [ "https://huggingface.co/baselabs/mammoth-1", "https://docs.baseten.co/reference/inference-api/system-one" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "elephant-1-local", "name": "Elephant-1 \u00b7 download", "provider": "Baseten / Base Labs", "providerId": "baselabs/elephant-1", "sourceUrl": "https://huggingface.co/baselabs/elephant-1", "officialUrl": "https://huggingface.co/baselabs/elephant-1", "discovered": "2026-10-23", "released": "2026-10-21", "context": 16384, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Run fixed-choice classification and scoring on your own hardware, with text or images.", "deployment": "checkpoint", "verification": "Download verified", "specialized": "decision", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "questionLimit": null, "contextNote": "The model card documents 16,384 maximum sequence length. Check the exact service if you use a hosted copy; a local limit does not guarantee that service\u2019s limits.", "accessNote": "MIT-labelled 9B weights and a pinned CUDA serving setup. The vision encoder is frozen. Check memory for all weights plus request data, and validate confidence on your own examples.", "pricingNote": "The hosted launch promotion does not pay for your local hardware. No measured local operating cost or independent comparison is established.", "checkpointRevision": "b3f8a93462fb1051a9de031e94a7ff380e254072", "capabilityNote": "Supply text or images and the choices you want the model to compare. Your application must carry out any action you choose from its answer.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "elephant-1-gateway", "name": "Elephant-1 \u00b7 Gateway", "provider": "Baseten / Base Labs", "providerId": "baseten/elephant-1", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "officialUrl": "https://ai-gateway.vercel.sh/v1/models", "discovered": "2026-10-23", "released": "2026-10-21", "context": 16384, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Choose from defined answers, route records or score a condition. Compare this model on your own examples before automating decisions.", "specialized": "decision", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "questionLimit": null, "identityLabel": "AI Gateway model", "verification": "Hosting offer verified", "documentedRates": { "input": 0.035, "output": 0, "cached": null, "label": "AI Gateway \u00b7 USD" }, "contextNote": "Gateway declares 16,384 tokens and text input. This service returns decisions rather than a generated written answer.", "capabilityNote": "The catalog documents choices, yes/no probabilities and ordered scores. Native image support and native question limits must not be assumed for this hosted offer.", "pricingNote": "The launch promotion has ended. Current standard input is $0.035 per million billed tokens, with no output-token charge. Native service-tier discounts do not automatically apply to Gateway.", "billingNote": "Use actual billed input tokens, including the questions. The current paid rate replaces the launch promotion.", "apiDocumentationUrls": [ "https://vercel.com/changelog/basetens-mammoth-1-and-elephant-1-now-available-on-ai-gateway-for-free" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "elephant-1", "name": "Elephant-1 \u00b7 Baseten API", "provider": "Baseten / Base Labs", "providerId": "decision-models/baseten-elephant-9b-latest", "sourceUrl": "https://docs.baseten.co/inference/model-apis/decision-models", "officialUrl": "https://docs.baseten.co/inference/model-apis/decision-models", "discovered": "2026-10-23", "released": "2026-10-21", "context": 16384, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Choose from defined answers, route records or score a condition. Compare this model on your own examples before automating decisions.", "specialized": "decision", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "questionLimit": null, "verification": "Native API verified", "documentedRates": { "input": 0.035, "output": 0, "cached": null, "label": "Native Baseten API \u00b7 USD" }, "contextNote": "Each question\u2019s full prompt must fit 16,384 tokens, including shared input, options, formatting and images. Over-limit requests are rejected, not shortened.", "capabilityNote": "Up to four images as base64 data URLs or supported cloud-storage HTTPS URLs. Questions are evaluated independently. Shared input is counted once per request, plus each question; confidence measures how concentrated the answer probabilities are.", "pricingNote": "The launch promotion has ended. Current standard input is $0.035 per million billed tokens, with no output-token charge. Native Flex is approximately 90% of default; priority defaults to twice the standard rate.", "billingNote": "Use actual billed input tokens, including the questions. The current paid rate replaces the launch promotion.", "accessNote": "Use the exact Baseten ID. Gateway\u2019s limits and modality tags differ. The publisher\u2019s evaluation claims are not independent scores here. Flex and priority are optional processing settings; faster completion is not independently measured here.", "apiDocumentationUrls": [ "https://huggingface.co/baselabs/elephant-1", "https://docs.baseten.co/reference/inference-api/system-one" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "recursive-flint-local", "name": "Recursive Flint \u00b7 download", "provider": "Recursive", "providerId": "RecursiveSI/flint", "sourceUrl": "https://huggingface.co/RecursiveSI/flint", "officialUrl": "https://huggingface.co/RecursiveSI/flint", "discovered": "2026-10-23", "released": null, "context": 262144, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Run a coding model on your own hardware to inspect a codebase, make changes and write tests.", "deployment": "checkpoint", "verification": "Download verified", "accessNote": "Apache 2.0 weights. The BF16 download is about 80 GB before runtime memory; the publisher supplies an SGLang kit for B300/H200 GPUs. Laptop use needs a suitable compressed format and engine.", "capabilityNote": "Text input and output, reasoning and tool calls. Preserve the returned reasoning history between turns. A verified adjustable effort menu is not available.", "contextNote": "Prompt, thinking and answer share 262,144 tokens. The suggested 32,768-token output budget is an example, not a universal maximum.", "pricingNote": "No local per-token price: include hardware and operating costs. Publisher benchmark and speed claims are not independent results.", "checkpointRevision": "2a247be0c0aec895ccb0ac7ca92229ec6df37288", "apiDocumentationUrls": [ "https://www.recursive.com/updates/introducing-recursive-flint" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "minimax-m3-1-flash-openrouter", "name": "MiniMax M3.1 Flash \u00b7 OpenRouter", "provider": "MiniMax", "providerId": "minimax/minimax-m3.1-flash", "sourceUrl": "https://openrouter.ai/minimax/minimax-m3.1-flash", "officialUrl": "https://openrouter.ai/minimax/minimax-m3.1-flash", "discovered": "2026-10-23", "released": "2026-10-19", "context": 1048576, "maxOutput": 524288, "modalities": [ "text", "image", "video" ], "tools": true, "note": "Use the stable model through a paid OpenRouter offer after the free route disappeared.", "identityLabel": "OpenRouter model ID", "verification": "Paid hosting offer verified", "defaultEffort": "max", "supportedEfforts": [ "low", "medium", "high", "xhigh", "max" ], "documentedRates": { "input": 0.15, "output": 0.6, "cached": 0.015, "label": "OpenRouter \u00b7 USD" }, "pricingNote": "Costs $0.15 input, $0.015 cached input and $0.60 output per million tokens. This is a paid offer, separate from the previous :free ID.", "accessNote": "Thinking is required. Verify application settings with the selected serving provider; the hosted offer does not establish download availability.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "deepseek-v4-1-flash-1017-cloudflare", "name": "DeepSeek V4.1 Flash 1017 \u00b7 Cloudflare", "provider": "DeepSeek / Cloudflare", "providerId": "@cf/deepseek-ai/deepseek-v4.1-flash-1017", "sourceUrl": "https://developers.cloudflare.com/workers-ai/models/deepseek-v4.1-flash-1017/", "officialUrl": "https://developers.cloudflare.com/workers-ai/platform/pricing/", "discovered": "2026-10-21", "released": null, "verification": "New hosting option", "context": 1048576, "maxOutput": null, "modalities": [ "text", "image" ], "tools": true, "note": "Use the October 17 DeepSeek model through Cloudflare Workers AI for text, images and tool-assisted work.", "supportedEfforts": [ "none", "low", "high", "max" ], "defaultEffort": "high", "effortScope": "Cloudflare Workers AI", "documentedRates": { "input": 0.3, "output": 1.2, "cached": 0.006, "label": "Cloudflare Workers AI \u00b7 USD" }, "pricingNote": "Lists $0.30 input, $0.006 cached input and $1.20 output per million tokens. Workers Paid or prepaid AI Gateway credit is required.", "contextNote": "Cloudflare advertises 1,048,576 context tokens. A maximum generated-output allowance is not confirmed; do not borrow another host\u2019s cap.", "capabilityNote": "The four documented thinking settings default to high. Compatibility names map to these settings: medium and auto to high, xhigh to max, minimal to none.", "accessNote": "Separate hosting for an existing revision, with its own billing and controls. No independent score or measured speed is inherited from another offer.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "scaledown-summarize-1-gateway", "name": "ScaleDown Summarize-1 \u00b7 Gateway", "provider": "ScaleDown", "providerId": "scaledown/scaledown-summarize-1", "identityLabel": "AI Gateway route", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "officialUrl": "https://vercel.com/ai-gateway/models/scaledown-summarize-1", "discovered": "2026-10-21", "released": null, "verification": "New hosting option", "specialized": "summarization", "context": 1000000, "maxOutput": 1000, "modalities": [ "text" ], "tools": false, "note": "Summarize a long document or transcript with instructions for language, length and focus.", "documentedRates": { "input": 0.05, "output": 0, "cached": null, "label": "Vercel AI Gateway \u00b7 USD" }, "contextNote": "Gateway lists one million context tokens and at most 1,000 output tokens. The separate OpenRouter offer lists a different output limit.", "pricingNote": "Gateway lists $0.05 per million input tokens and zero output-token charge. No cache rate is verified.", "accessNote": "This is a new Gateway offer for an existing specialized model. Use the full service ID; native API limits and document-processing features may differ.", "dateNote": "Gateway dates this offer October 20. This does not establish a new underlying checkpoint.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "scaledown-extract-1-gateway", "name": "ScaleDown Extract-1 \u00b7 Gateway", "provider": "ScaleDown", "providerId": "scaledown/scaledown-extract-1", "identityLabel": "AI Gateway route", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "officialUrl": "https://vercel.com/ai-gateway/models/scaledown-extract-1", "discovered": "2026-10-21", "released": null, "verification": "New hosting option", "specialized": "extraction", "context": 1000000, "maxOutput": 1000000, "modalities": [ "text" ], "tools": false, "note": "Extract fields from a document using the structure you define, such as invoice amounts or contract parties.", "documentedRates": { "input": 0.05, "output": 0, "cached": null, "label": "Vercel AI Gateway \u00b7 USD" }, "contextNote": "Gateway declares one million context and output tokens. The native extractor separately limits distinct values per entity type; do not interpret the hosted output allowance as a promise of one million useful extracted tokens.", "pricingNote": "Gateway lists $0.05 per million input tokens and zero output-token charge. No cache rate is verified.", "accessNote": "This is a new Gateway offer for an existing specialized model. Use the full service ID; native API limits and document-processing features may differ.", "dateNote": "Gateway dates this offer October 20. This does not establish a new underlying checkpoint.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "xing4-0-29b-intent", "name": "Xing4.0 29B Intent", "provider": "XingChen / China Telecom", "providerId": "XingChen-AGI/Xing4.0-29B-Intent", "sourceUrl": "https://huggingface.co/XingChen-AGI/Xing4.0-29B-Intent", "officialUrl": "https://huggingface.co/XingChen-AGI/Xing4.0-29B-Intent", "discovered": "2026-10-21", "released": null, "verification": "Download verified", "deployment": "checkpoint", "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": false, "note": "Turn a request such as \u201cfind flights tomorrow\u201d into a category and fields your application can use. Supply your categories, field descriptions and text or conversation.", "specialized": "extraction", "capabilityNote": "Returns JSON containing categories and extracted arguments. Chinese and English examples cover intent, entities, relations and events. This is structured extraction, not a fixed-choice probability model.", "contextNote": "The checkpoint configuration lists 262,144 positions, but the specialized instructions do not establish a validated task input or output limit. Check your serving setup and real documents.", "accessNote": "Apache 2.0 weights and companion software. Roughly 29B total parameters with about 4B active per token still require storage for all weights. The repository was created October 15 and updated October 20; its exact release day remains unverified. SGLang and vLLM require the publisher\u2019s separate plugin; an unmodified installation does not recognize this model.", "pricingNote": "No hosted price verified. FP8 and GGUF are formats of this model, not additional base releases.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "qev-0-8b-labqoat", "name": "qev 0.8B \u00b7 Labqoat", "provider": "Labqoat", "providerId": "labqoat/qev-0.8b", "sourceUrl": "https://huggingface.co/labqoat/qev-0.8b", "officialUrl": "https://huggingface.co/labqoat/qev-0.8b", "discovered": "2026-10-21", "released": null, "verification": "Download verified", "deployment": "checkpoint", "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "note": "Sort messages or select a next action locally using answer choices you define. This is Labqoat\u2019s model; similarly named Qev downloads from other authors are separate.", "specialized": "decision", "outputType": "Yes/no probabilities, choices and ordered scores; no generated explanation.", "questionLimit": null, "contextNote": "A dependable task-specific input limit is not established: the decision configuration allows one million state tokens while the inherited model configuration lists 262,144 positions. Long-input accuracy needs separate testing.", "capabilityNote": "The current SG50 revision adds image input through its supplied runtime. Images are resized to at most 448\u00d7448 pixels; image support uses the base vision encoder without specialized image training. Use separate evaluation for each question on image or closely related requests.", "accessNote": "Apache 2.0-labelled weights with CUDA, CPU, MLX and GGUF paths. The publisher flags mixed training-data terms, weak scoring, difficult long inputs and unreliable abstention. Its fitted probabilities need calibration on your own examples. Pin the revision: the current artifact differs from the earlier text-only release.", "pricingNote": "No hosted rate. Publisher tests and speed figures are not independent measurements. Repository first created October 20; pin the exact revision because tokenizer and weights can change.", "checkpointRevision": "9a095f35fe0e27a888cfed9eb1517718dc9dc59e", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "ornith-1-5-122b", "name": "Ornith 1.5 122B", "provider": "Ornith", "providerId": "ornith-ai/Ornith-1.5-122B", "sourceUrl": "https://huggingface.co/ornith-ai/Ornith-1.5-122B", "officialUrl": "https://huggingface.co/ornith-ai/Ornith-1.5-122B", "discovered": "2026-10-21", "released": null, "verification": "Download verified", "deployment": "checkpoint", "context": 262144, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Run a large reasoning model yourself for coding and tool-assisted work. Choose the weight format to fit your hardware before planning a deployment.", "contextNote": "The documented standard window is 262,144 tokens. Extending it toward one million requires explicit scaling and can reduce quality on shorter inputs.", "capabilityNote": "Thinking is enabled by default; the serving guide supports tool calls. No adjustable effort menu or independently matched result is verified here.", "accessNote": "The publisher labels the weights MIT. Roughly 250 GB of raw BF16 weights require suitable hardware before runtime memory; compressed formats need their own evaluation. Use the exact runtime and weight format when comparing memory requirements.", "pricingNote": "A download, with no verified hosted token price. Publisher evaluations do not establish an independent ranking.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "spark-x2-6-4b", "name": "Spark-X2.6 4B", "provider": "XHToken", "providerId": "XHToken/Spark-X2.6-4B", "sourceUrl": "https://huggingface.co/XHToken/Spark-X2.6-4B", "officialUrl": "https://huggingface.co/XHToken/Spark-X2.6-4B", "discovered": "2026-10-21", "released": null, "verification": "Download verified", "deployment": "checkpoint", "context": 1048576, "maxOutput": null, "modalities": [ "text" ], "tools": true, "note": "Run a compact model locally for coding, text tasks and tool-assisted workflows. It builds on Spark X2.5 with additional training.", "supportedEfforts": [ "none", "low", "high" ], "defaultEffort": "high", "capabilityNote": "Use the low or high thinking setting, or turn thinking off. Your application still needs to execute tools and manage permissions.", "contextNote": "The configuration and serving guide allow 1,048,576 combined tokens, subject to available memory. The example\u2019s 131,072 output budget is not a verified universal output limit.", "accessNote": "Apache 2.0 weights. Use a compatible Spark runtime or plugin; local hardware and memory determine practical capacity. The repository existed October 19 and was updated October 20; an exact launch day is not established. The publisher now links a community browser demo using the INT4 weights. It was tested by its contributor on Windows with an NVIDIA GPU; that does not establish compatibility or speed on every browser or device.", "pricingNote": "No hosted token price verified. Budget for your own hardware and operation.", "effortScope": "chat-template", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "cloudflare-clairic", "name": "Clairic \u00b7 Cloudflare", "provider": "Cloudflare", "providerId": "@cf/cloudflare/clairic", "discovered": "2026-10-20", "released": null, "verification": "Preview API verified", "specialized": "extraction", "context": 65536, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "sourceUrl": "https://developers.cloudflare.com/workers-ai/models/clairic/", "officialUrl": "https://developers.cloudflare.com/workers-ai/concepts/extraction-models/", "apiDocumentationUrls": [ "https://developers.cloudflare.com/workers-ai/platform/pricing/" ], "documentedRates": { "input": 0.1, "output": null, "cached": null, "label": "Cloudflare Workers AI \u00b7 USD" }, "note": "Turn an invoice or screenshot into useful fields, such as an amount, due date or estimated payment time. Missing values and uncertain predictions need different handling.", "capabilityNote": "Returns text spans, numbers, dates, lists, estimates and forecasts with uncertainty information. A valid format or high confidence does not guarantee a correct value. Estimates can still be returned when evidence is sparse; check important results before acting.", "contextNote": "The model page advertises 65,536 context tokens, while its schema splits fields across 8,192-token processing passes. Check whether the response reports truncated input, and use its usage totals for billing. Up to four embedded PNG/JPEG/WebP images: 4 MiB each, 8 MiB total decoded and a 16 MiB request body; no remote image URLs.", "pricingNote": "Cloudflare lists $0.10 per million input tokens. A separate output rate and cached-input policy are not verified. Several processing passes can be needed; use reported input_tokens rather than the original text length.", "accessNote": "Available as a Workers AI preview. It extracts or predicts values, which makes it different from a model that only chooses among your predefined answers.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "drex-max", "name": "Drex Max", "provider": "Nace.AI", "providerId": "drex-max", "discovered": "2026-10-20", "released": "2026-10-13", "verification": "Native model verified", "specialized": "decision", "context": 262144, "maxOutput": null, "questionLimit": 512, "modalities": [ "text", "image" ], "tools": false, "sourceUrl": "https://drex.nace.ai/docs/drex/guides/drex-max", "officialUrl": "https://drex.nace.ai/docs/drex/reference/models", "apiDocumentationUrls": [ "https://drex.nace.ai/docs/drex/reference/limits", "https://drex.nace.ai/docs/drex/reference/pricing" ], "documentedRates": { "input": 0.03, "output": 0, "cached": null, "label": "Native Drex API \u00b7 USD" }, "note": "Check text and images against up to 512 questions\u2014for example, identify a damaged parcel and choose how to handle the claim.", "outputType": "Yes/no probabilities, choices and ordered scores; no written explanation.", "capabilityNote": "Reads text and images and returns probabilities for choices, yes/no questions and ratings. Up to 512 questions are allowed; that limit does not guarantee completion within the service timeout.", "contextNote": "Original Drex Max permits up to 512 questions, 255 options per question and 8 images. Its 262,144-token limit applies to shared information plus each question; long questions also share an aggregate budget. It can shorten long state before inference. Version v1.1 instead counts all questions in one whole-request token budget.", "pricingNote": "Costs $0.03 per million billed input tokens, including image tokens. Output and failed requests are not charged. Use the successful response\u2019s input usage to calculate the bill.", "billingNote": "Image tokens already belong to billed input. Output and failed requests are not charged; use successful response usage totals.", "accessNote": "Select drex-max explicitly. The drex-latest alias and omitted model still choose v2.0. Larger accepted requests can time out; fewer questions may be needed for long inputs. Check the available-model list with your actual API key; availability can differ by environment.", "dateNote": "The provider\u2019s model table dates this version October 13; first verified in this comparison October 20.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "mercury-decide", "name": "Mercury Decide", "provider": "Inception", "providerId": "mercury-decide", "discovered": "2026-10-20", "released": null, "verification": "Native API verified", "specialized": "decision", "context": null, "maxOutput": null, "questionLimit": null, "modalities": [ "text", "image" ], "tools": false, "sourceUrl": "https://docs.inceptionlabs.ai/capabilities/decisions", "officialUrl": "https://www.inceptionlabs.ai/blog/introducing-mercury-3", "apiDocumentationUrls": [ "https://docs.inceptionlabs.ai/get-started/models", "https://docs.inceptionlabs.ai/capabilities/extract" ], "documentedRates": { "input": 0.02, "output": 0, "cached": 0, "label": "Native launch discount \u00b7 USD" }, "note": "Route a support request, choose a category or score urgency. Send shared information and named questions; the service returns probabilities and predefined answers.", "outputType": "Choices, yes/no probabilities, ordered scores, or exact text spans with character positions.", "capabilityNote": "Questions are evaluated independently against shared information. Choice questions allow 1\u2013255 options and scores 1\u201310 levels. Images require base64 data, not remote URLs. Up to 16 images, 10 MiB each, must fit within the 20 MiB total JSON request. Extraction can return matching text and its character positions for 1\u201326 fields per question, with confidence and an indication that a field was not stated. It selects from the input rather than writing new text. Requests using extraction cannot include images; use a separate request for image decisions.", "contextNote": "A context-token limit and maximum question count are not confirmed. The 20 MiB body allowance is not a token limit. Image encoding adds roughly one third to file size.", "pricingNote": "The launch input price is $0.02 per million billed tokens, half the regular $0.04. Output is not charged. The model table lists cached input at zero; cache eligibility and a promotion end date are not verified.", "billingNote": "Launch input price only; use the returned billed input totals. Output is free. The regular input rate is $0.04/M, with no verified promotion end date.", "accessNote": "Use /v1/decisions with mercury-decide. Chat completions, streaming and tool calling are unsupported. Validate probability thresholds on your own examples.", "dateNote": "Mentioned in the October 19 Mercury 3 announcement; an exact separate launch day is not verified.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "minimax-m3-1-flash", "name": "MiniMax M3.1 Flash \u00b7 native API", "provider": "MiniMax", "providerId": "MiniMax-M3.1-Flash", "discovered": "2026-10-20", "released": "2026-10-19", "verification": "Stable release verified", "context": 1000000, "maxOutput": 524288, "modalities": [ "text", "image", "video" ], "tools": true, "supportedEfforts": [ "low", "medium", "high", "xhigh", "max" ], "defaultEffort": "max", "sourceUrl": "https://platform.minimax.io/docs/api-reference/text-openai-api", "officialUrl": "https://www.minimax.io/blog/minimax-m31f", "apiDocumentationUrls": [ "https://platform.minimax.io/docs/pricing/overview", "https://platform.minimax.io/docs/api-reference/text-chat-openai", "https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash", "https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash/blob/main/LICENSE" ], "documentedRates": { "input": 0.15, "output": 0.6, "cached": 0.015, "label": "MiniMax direct API \u00b7 USD" }, "note": "Use text, screenshots and video in coding or multi-step tasks. This is the stable October release, separate from the earlier subscription preview.", "capabilityNote": "Thinking is always on; requests to turn it off are rejected. Tools can run between thinking steps, but your application must execute them and control permissions. Sampling controls are currently fixed: temperature stays 1 and top_p stays 0.95; supplied values do not change them.", "contextNote": "The API documents a 1,000,000-token context and up to 524,288 generated tokens. Thinking and the final answer share the output budget.", "pricingNote": "The direct API charges $0.15 input, $0.015 cached input and $0.60 output per million tokens. These flat rates are separate from M Plan subscription allowances. Priority processing is not supported for this model (the service_tier setting). Downloaded weights use the MiniMax Community License: commercial use requires attribution and a one-time notice; products/services above $20M yearly revenue require prior written authorization. Read the full license for conditions.", "accessNote": "Also downloadable as MiniMaxAI/MiniMax-M3.1-Flash. The local template supports low, medium, high, xhigh and max; its recommended sampling settings do not change the hosted API\u2019s fixed controls. Space Bunny was an early version, not an identical tested checkpoint.", "dateNote": "The provider dates the stable release October 19. Performance and speed claims in its announcement are vendor evidence, not independent results here.", "checkpointRevision": "262ae6830934f3512211081b8a3f16425b93a239", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "deepseek-v4-1-flash-1017-fast", "name": "DeepSeek V4.1 Flash 1017 Fast \u00b7 Gateway", "provider": "DeepSeek / Baseten", "providerId": "deepseek/deepseek-v4.1-flash-1017-fast", "identityLabel": "AI Gateway route", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "officialUrl": "https://vercel.com/ai-gateway/models/deepseek-v4.1-flash-1017-fast", "discovered": "2026-10-19", "released": null, "verification": "New hosting option", "note": "Run the October 17 Flash model through Baseten on AI Gateway. This option supports long inputs and cached prompts. Test response times on your own workload before paying more for the Fast offer.", "context": 1048576, "maxOutput": 262144, "modalities": [ "text", "image" ], "tools": true, "documentedRates": { "input": 0.6, "output": 2.4, "cached": 0.014, "label": "AI Gateway / Baseten offer \u00b7 USD" }, "pricingNote": "Gateway lists $0.60 input, $0.014 cached input and $2.40 output per million tokens. Automatic caching reuses matching prompt prefixes. These are prices for this hosting option; DeepSeek\u2019s direct API uses a different peak/off-peak schedule.", "contextNote": "Gateway lists 1,048,576 context tokens and 262,144 maximum output tokens. The older unversioned Fast offer lists 32,768 maximum output tokens; it is a separate offer.", "capabilityNote": "Use text and images with tools or structured answers. Baseten markets this option for faster output, but we do not yet have a same-task test showing its response time or total bill.", "gatewayEffortOptions": [ "none", "low", "medium", "high" ], "capabilitySourceUrl": "https://ai-gateway.vercel.sh/v1/models", "dateNote": "Gateway dates the underlying model October 17. This hosting option was first verified here October 19; it is not another base-model release.", "accessNote": "Use the full versioned ID to select this offer. Its listed thinking controls differ from the ordinary 1017 OpenRouter offer, which includes max; do not assume the controls produce equivalent results.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "granite-switch-4-1-30b", "name": "Granite Switch 4.1.1 30B", "provider": "IBM", "providerId": "ibm-granite/granite-switch-4.1-30b", "sourceUrl": "https://huggingface.co/ibm-granite/granite-switch-4.1-30b", "officialUrl": "https://huggingface.co/ibm-granite/granite-switch-4.1-30b", "note": "Run chat, classification and speech transcription on one server. Separate built-in components handle general decisions and harmful-content checks, so you can reuse the same deployment for several steps of a workflow.", "discovered": "2026-10-17", "released": "2026-10-14", "context": 131072, "maxOutput": null, "modalities": [ "text", "audio" ], "tools": null, "deployment": "checkpoint", "specialized": "decision", "questionLimit": null, "outputType": "Chat text, probabilities over supplied options, safety checks and speech transcripts", "license": "Apache 2.0", "capabilityNote": "Revision 4.1.1 updates the decision adapter and speech recognizer while keeping the base language model unchanged. Pin the package revision when comparing results; improvements are publisher reports, not new independent measurements.", "contextNote": "The base model has a 128K context window. The exact packaged maximum output and question count are not confirmed. The documented vLLM probability lookup supports 128 option-token IDs per call; more options require repeated calls with the same prompt.", "pricingNote": "Apache 2.0 permits downloading and running the weights. There is no verified per-token hosted price here; budget for hardware, memory and serving software.", "accessNote": "Model cards date these composed checkpoints October 14; IBM\u2019s launch article is October 15. They combine an existing Granite 4.1 base with decision and Guardian adapters plus a speech encoder. These are distinct from the earlier preview bundles. The cards date revision 4.1.1 October 27; IBM announced it October 29. Use granite-switch 0.1.1 or newer. The announcement and model cards differ on speech language counts, so confirm the languages in your exact package.", "apiDocumentationUrls": [ "https://huggingface.co/blog/ibm-granite/granite-switch-4-1", "https://huggingface.co/blog/ibm-granite/granite-switch-4-1-1" ], "identityLabel": "downloaded model", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "granite-switch-4-1-8b", "name": "Granite Switch 4.1.1 8B", "provider": "IBM", "providerId": "ibm-granite/granite-switch-4.1-8b", "sourceUrl": "https://huggingface.co/ibm-granite/granite-switch-4.1-8b", "officialUrl": "https://huggingface.co/ibm-granite/granite-switch-4.1-8b", "note": "Run chat, classification and speech transcription on one server. Separate built-in components handle general decisions and harmful-content checks, so you can reuse the same deployment for several steps of a workflow.", "discovered": "2026-10-17", "released": "2026-10-14", "context": 131072, "maxOutput": null, "modalities": [ "text", "audio" ], "tools": null, "deployment": "checkpoint", "specialized": "decision", "questionLimit": null, "outputType": "Chat text, probabilities over supplied options, safety checks and speech transcripts", "license": "Apache 2.0", "capabilityNote": "Revision 4.1.1 updates the decision adapter and speech recognizer while keeping the base language model unchanged. Pin the package revision when comparing results; improvements are publisher reports, not new independent measurements.", "contextNote": "The base model has a 128K context window. The exact packaged maximum output and question count are not confirmed. The documented vLLM probability lookup supports 128 option-token IDs per call; more options require repeated calls with the same prompt.", "pricingNote": "Apache 2.0 permits downloading and running the weights. There is no verified per-token hosted price here; budget for hardware, memory and serving software.", "accessNote": "Model cards date these composed checkpoints October 14; IBM\u2019s launch article is October 15. They combine an existing Granite 4.1 base with decision and Guardian adapters plus a speech encoder. These are distinct from the earlier preview bundles. The cards date revision 4.1.1 October 27; IBM announced it October 29. Use granite-switch 0.1.1 or newer. The announcement and model cards differ on speech language counts, so confirm the languages in your exact package.", "apiDocumentationUrls": [ "https://huggingface.co/blog/ibm-granite/granite-switch-4-1", "https://huggingface.co/blog/ibm-granite/granite-switch-4-1-1" ], "identityLabel": "downloaded model", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "elorian-glimpse", "name": "Elorian Glimpse", "provider": "Elorian", "providerId": "elorian/glimpse", "sourceUrl": "https://elorian.ai/blog/introducing-glimpse", "officialUrl": "https://elorian.ai/blog/introducing-glimpse", "note": "Analyze charts, follow objects in video or count items in images. Glimpse is a research preview for visual questions; its maker\u2019s comparisons use tools and settings that differ across models, so they are not a like-for-like quality ranking here.", "discovered": "2026-10-17", "released": "2026-10-16", "context": 220000, "maxOutput": 4096, "modalities": [ "text", "image", "video" ], "tools": null, "identityLabel": "AI Gateway route", "verification": "Research preview", "documentedRates": { "input": 0.6, "output": 3, "cached": 0.06, "label": "AI Gateway offer \u00b7 USD" }, "pricingNote": "Gateway lists $0.60 input, $0.06 cached input and $3 output per million tokens. Elorian\u2019s launch also quotes $0.60 input and $3 output; native caching terms are not established.", "contextNote": "Gateway lists 220,000 context tokens and 4,096 maximum output tokens. These limits describe that hosted offer.", "supportedEfforts": [ "low", "medium", "xhigh" ], "effortScope": "AI Gateway offer", "capabilityNote": "Gateway accepts text, images and video and lists low, medium and xhigh thinking. Elorian\u2019s published evaluations use a separate max setting and built-in visual tools; do not assume those tools or max effort are exposed by Gateway.", "accessNote": "Elorian offers a web app and API. Gateway\u2019s exact ID is elorian/glimpse; the native API ID is not yet verified.", "apiDocumentationUrls": [ "https://ai-gateway.vercel.sh/v1/models" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "perplexity-decider-flash-v1-2", "name": "Perplexity Decider Flash v1.2", "provider": "Perplexity", "providerId": "pplx-decider-flash-v1.2", "sourceUrl": "https://docs.perplexity.ai/docs/decisions/quickstart", "officialUrl": "https://docs.perplexity.ai/docs/decisions/quickstart", "note": "A lower-priced option for large volumes of clear decisions, such as labeling messages or routing tickets. Test it against examples you have already labeled before choosing it on price.", "discovered": "2026-10-17", "released": "2026-10-16", "context": 262144, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "dateNote": "Gateway dates its new offers October 16; the native changelog records the release in October without a day.", "specialized": "decision", "questionLimit": 128, "outputType": "Yes/no probabilities, choices and ordered scores", "documentedRates": { "input": 0.01, "output": 0, "cached": null, "label": "Direct Decisions API \u00b7 USD" }, "contextNote": "The native limit is fewer than 256K tokens for the shared information, images and each question. Large multi-question requests can be split. Up to 128 questions, 255 choice options, 10 score levels and a 32 MiB request body. Gateway\u2019s 262,144 declaration is separate.", "capabilityNote": "Each question is answered separately, even when several are sent in one request. Start with one question per request; other questions do not share one inference pass.", "pricingNote": "Costs $0.01 per million input tokens, with free output and no per-request fee. Shared information is sent and billed again for every question. Three questions about 10,000 shared tokens bill roughly 30,000 input tokens, plus the questions.", "billingNote": "Repeat the shared input for every question. For three or more questions, regular Decider can cost less if they fit in one group.", "accessNote": "Use pplx-decider-flash-v1.2 for the current version or pplx-decider-flash-latest to follow future updates. Revalidate your probability thresholds after a version change. Gateway also lists perplexity/pplx-decider-flash-v1.2. Its zero output-limit placeholder is not a verified limit on answers. The native default is 10 requests per second per organization shared across both models, plus 4 million tokens per minute and 240 million per hour for this model. Higher limits require contacting Perplexity.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "perplexity-decider-v1-2", "name": "Perplexity Decider v1.2", "provider": "Perplexity", "providerId": "pplx-decider-v1.2", "sourceUrl": "https://docs.perplexity.ai/docs/decisions/quickstart", "officialUrl": "https://docs.perplexity.ai/docs/decisions/quickstart", "note": "A decision model for checking claims, grading records against a rubric or routing difficult requests. It returns probabilities and predefined choices, rather than a written answer.", "discovered": "2026-10-17", "released": "2026-10-16", "context": 262144, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "dateNote": "Gateway dates its new offers October 16; the native changelog records the release in October without a day.", "specialized": "decision", "questionLimit": 128, "outputType": "Yes/no probabilities, choices and ordered scores", "documentedRates": { "input": 0.02, "output": 0, "cached": null, "label": "Direct Decisions API \u00b7 USD" }, "contextNote": "The native limit is fewer than 256K tokens for the shared information, images and each question. Large multi-question requests can be split. Up to 128 questions, 255 choice options, 10 score levels and a 32 MiB request body. Gateway\u2019s 262,144 declaration is separate.", "capabilityNote": "Related questions can be answered together and share one input charge. Each question sees the others, which can change the answer. Score independent competing documents or candidates in separate requests. Confidence is calculated from the answer probabilities, not a separate guarantee of correctness.", "pricingNote": "Native input costs $0.02 per million tokens; answers are free and there is no per-request fee. Shared information is billed once within a group. If the service splits a large request into groups, it counts that information again.", "billingNote": "Use actual billed input tokens. Grouping can reduce repeated input, but changes answers; hold question grouping fixed when comparing models.", "accessNote": "Use pplx-decider-v1.2 for the current version or pplx-decider-latest to follow future updates. Revalidate your probability thresholds after a version change. Gateway also lists perplexity/pplx-decider-v1.2. Its zero output-limit placeholder is not a verified limit on answers. The native default is 10 requests per second per organization shared across both models, plus 15 million tokens per minute and 900 million per hour for this model. Higher limits require contacting Perplexity.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "phonellm-mini-alpha-2", "name": "PhoneLLM Mini Alpha 2 \u00b7 Cloudflare", "provider": "Daily / Cloudflare", "providerId": "@cf/daily-co/phonellm-mini-alpha-2", "sourceUrl": "https://developers.cloudflare.com/workers-ai/models/phonellm-mini-alpha-2/", "officialUrl": "https://developers.cloudflare.com/workers-ai/models/phonellm-mini-alpha-2/", "discovered": "2026-10-16", "released": null, "context": 262140, "maxOutput": null, "modalities": [ "text" ], "tools": true, "documentedRates": { "input": 0.06, "output": 0.24, "cached": null, "label": "Cloudflare Workers AI \u00b7 USD" }, "note": "Generate replies and call tools inside a voice assistant. This Cloudflare offer provides the text model; speech recognition and spoken output need their own components.", "capabilityNote": "Cloudflare describes a model tuned for voice-agent conversations with thinking disabled. Its text-generation endpoint supports tools. Generic audio fields in the API schema do not establish built-in speech support for this model.", "contextNote": "Cloudflare lists exactly 262,140 context tokens. The maximum generated response and a launch date are not verified.", "accessNote": "Connect through a Cloudflare Workers AI binding or its REST API. A separate speech-to-text and text-to-speech service can turn its text replies into a spoken conversation.", "pricingNote": "Cloudflare quotes $0.06 input and $0.24 output per million tokens. Speech services and other application costs are separate.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "pareto-26-10", "name": "Pareto 26.10 \u00b7 native stable", "provider": "Unbiased", "providerId": "pareto-26.10", "sourceUrl": "https://unbiased.ai/blog/pareto-26-10/", "officialUrl": "https://unbiased.ai/blog/pareto-26-10/", "note": "Send a request to a service that combines several models into one answer. This is the stable October release; the older OpenRouter preview page is a separate offer.", "discovered": "2026-10-16", "released": "2026-10-15", "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": null, "verification": "Native stable composite service", "documentedRates": { "input": 0.8, "output": 3.2, "cached": 0.03, "label": "Unbiased native API \u00b7 USD" }, "contextNote": "The provider announces a one-million-token context. Exact token boundary and maximum generated output are not verified here.", "accessNote": "The native pareto alias now selects stable 26.10. Use pareto-26.10 to choose it explicitly, or pareto-26.9 to retain the older version. Accounts are reviewed manually. Data-retention terms depend on your direct plan. Unbiased\u2019s zero-retention promise applies to API traffic explicitly designated for it. Chat and Personal subscriptions are outside that tier; saved conversations and projects can remain until deleted. Check the plan\u2019s data policy before sending private material.", "capabilityNote": "The displayed provider evaluations are preliminary October 1 preview runs, not new measurements of the stable release. No independent score is assigned here.", "pricingNote": "The provider kept preview pricing at graduation: $0.80 input, $0.03 cached input and $3.20 output per million tokens. Personal subscription allowances doubled; subscriptions and API token prices are different billing plans.", "catalogOffer": { "id": "unbiased/pareto-26.10", "input": 0.8, "output": 3.2, "cached": 0.03, "checked": "2026-11-03", "observed": "2026-11-03", "note": "Current OpenRouter rates are shown separately. Use the prices for the service your application calls.", "sourceUrl": "https://openrouter.ai/unbiased/pareto-26.10" }, "apiDocumentationUrls": [ "https://unbiased.ai/terms" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "perplexity-nemotron-3-5-super", "name": "Nemotron 3.5 Super \u00b7 Perplexity", "provider": "NVIDIA / Perplexity", "providerId": "perplexity/nemotron-3.5-super-120b-a12b", "sourceUrl": "https://docs.perplexity.ai/docs/agent-api/models", "officialUrl": "https://docs.perplexity.ai/docs/agent-api/models", "note": "Use this text-only Nemotron reasoning offer through Perplexity\u2019s Agent API. It replaces the older Ultra offer in that service\u2019s model list.", "discovered": "2026-10-16", "released": null, "context": 1048576, "maxOutput": 131072, "modalities": [ "text" ], "tools": null, "documentedRates": { "input": 0.2, "output": 0.5, "cached": 0.04, "label": "Perplexity Agent API \u00b7 USD" }, "contextNote": "The documented limit is 1,048,576 combined context tokens, with up to 131,072 generated tokens.", "capabilityNote": "This offer accepts text only. The similarly named Super VL model\u2019s image and video features are not promised by this endpoint.", "accessNote": "The older perplexity/nemotron-3-ultra-550b-a55b no longer appears in the Agent and Orchestrator model documentation. This does not establish retirement at other providers.", "pricingNote": "The native table lists $0.20 input, $0.04 cached input and $0.50 output per million tokens. No special service tier is listed for this model.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "liquid-d1-ultra", "name": "Liquid d1-ultra", "provider": "Liquid AI", "providerId": "d1-ultra", "sourceUrl": "https://www.liquid.ai/blog/d1-ultra", "officialUrl": "https://www.liquid.ai/blog/d1-ultra", "note": "Check a photo against a damage claim, inspect a component or route a difficult support case. Supply the possible answers; the model returns probabilities rather than writing a reply.", "discovered": "2026-10-16", "released": "2026-10-15", "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "specialized": "decision", "questionLimit": null, "outputType": "Yes/no probabilities, choices and ordered scores", "documentedRates": { "input": 0.042, "output": 0, "cached": null, "label": "Direct Liquid API \u00b7 USD" }, "capabilityNote": "Accepts text and images for yes/no, choice and score questions. The launch compares it with d1 and Jev, but those are the developer\u2019s tests, not independent measurements in this ranking.", "contextNote": "Native context and question limits are not yet verified. Gateway separately lists liquid/d1-ultra at 32,768 tokens; that does not establish the direct API limit.", "pricingNote": "Each question is billed as its own prompt, including all images. A 1024\u00d71024 image counts as 1,536 input tokens per question. Use the sum of billed tokens across questions in the calculator.", "billingNote": "Input costs $0.042 per million tokens. Repeating an image across questions repeats its input charge; no output tokens are generated.", "accessNote": "Use d1-ultra with Liquid\u2019s decision endpoint. Gateway also lists a separate offer; check that service\u2019s limits and billing before switching.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jev-1-14-0", "name": "Jev 1.14.0 \u00b7 Gateway", "provider": "TypeSafe AI", "providerId": "typesafe-ai/jev-1.14.0", "sourceUrl": "https://docs.typesafe.ai/models", "officialUrl": "https://docs.typesafe.ai/models", "note": "Choose a fixed Jev version when you need repeatable ticket routing, labels or scores. Retest your confidence threshold before changing versions.", "discovered": "2026-10-16", "released": "2026-10-14", "context": 32000, "maxOutput": null, "modalities": [ "text" ], "tools": false, "identityLabel": "AI Gateway route", "specialized": "decision", "questionLimit": null, "outputType": "Choices, ordered scores and yes/no probabilities", "documentedRates": { "input": 0.042, "output": 0, "cached": null, "label": "AI Gateway offer \u00b7 USD" }, "contextNote": "Gateway lists 32,000 tokens. Native Jev documents 32,000 for shared information plus one question and 64,000 for the whole request. Its optional longer-context header is not assumed to work through Gateway.", "capabilityNote": "The native model table describes text input; Gateway also tags vision. Image support for this exact route remains unresolved, so this page does not promise it.", "accessNote": "Native jev-latest and jev-preview now both select 1.14.0; there is no separate preview build at this check. Request an exact version for reproducible tests. Gateway uses the separate ID shown here.", "billingNote": "The estimate uses Gateway\u2019s $0.042 per million input tokens. Output-token price is zero; caching and other fees are unverified.", "pricingNote": "The same input quote does not prove equal accuracy. Compare both versions on your own labeled examples.", "apiDocumentationUrls": [ "https://ai-gateway.vercel.sh/v1/models" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jev-1-13-0", "name": "Jev 1.13.0 \u00b7 Gateway", "provider": "TypeSafe AI", "providerId": "typesafe-ai/jev-1.13.0", "sourceUrl": "https://docs.typesafe.ai/models", "officialUrl": "https://docs.typesafe.ai/models", "note": "Choose a fixed Jev version when you need repeatable ticket routing, labels or scores. Retest your confidence threshold before changing versions.", "discovered": "2026-10-16", "released": null, "context": 32000, "maxOutput": null, "modalities": [ "text" ], "tools": false, "identityLabel": "AI Gateway route", "specialized": "decision", "questionLimit": null, "outputType": "Choices, ordered scores and yes/no probabilities", "documentedRates": { "input": 0.042, "output": 0, "cached": null, "label": "AI Gateway offer \u00b7 USD" }, "contextNote": "Gateway lists 32,000 tokens. Native Jev documents 32,000 for shared information plus one question and 64,000 for the whole request. Its optional longer-context header is not assumed to work through Gateway.", "capabilityNote": "The native model table describes text input; Gateway also tags vision. Image support for this exact route remains unresolved, so this page does not promise it.", "accessNote": "Native jev-latest and jev-preview now both select 1.14.0; there is no separate preview build at this check. Request an exact version for reproducible tests. Gateway uses the separate ID shown here.", "billingNote": "The estimate uses Gateway\u2019s $0.042 per million input tokens. Output-token price is zero; caching and other fees are unverified.", "pricingNote": "The same input quote does not prove equal accuracy. Compare both versions on your own labeled examples.", "apiDocumentationUrls": [ "https://ai-gateway.vercel.sh/v1/models" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "apertus-1-5-70b", "name": "Apertus 1.5 70B", "provider": "Swiss AI", "providerId": "swiss-ai/Apertus-v1.5-70B", "sourceUrl": "https://huggingface.co/blog/swiss-ai/apertus-v15-release-notes", "officialUrl": "https://huggingface.co/blog/swiss-ai/apertus-v15-release-notes", "note": "An openly documented multilingual model for text, images and experimental audio input. Consider it when you want to run or adapt a model yourself and inspect how it was trained.", "discovered": "2026-10-16", "released": "2026-07-24", "context": 262144, "maxOutput": null, "modalities": [ "text", "image", "audio" ], "tools": true, "deployment": "checkpoint", "identityLabel": "checkpoint", "capabilityNote": "Produces text and supports optional thinking and tools. Audio understanding remains experimental. The October 15 article explains an earlier July 24 release; it is not a new October checkpoint.", "accessNote": "The project publishes Apache 2.0 weights and training details. Choose the 8B or 70B download to match your available hardware, and follow the project\u2019s setup guide.", "pricingNote": "No hosted token quote is verified for this exact profile. Local hardware and operating costs are separate.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "apertus-1-5-8b", "name": "Apertus 1.5 8B", "provider": "Swiss AI", "providerId": "swiss-ai/Apertus-v1.5-8B", "sourceUrl": "https://huggingface.co/blog/swiss-ai/apertus-v15-release-notes", "officialUrl": "https://huggingface.co/blog/swiss-ai/apertus-v15-release-notes", "note": "An openly documented multilingual model for text, images and experimental audio input. Consider it when you want to run or adapt a model yourself and inspect how it was trained.", "discovered": "2026-10-16", "released": "2026-07-24", "context": 262144, "maxOutput": null, "modalities": [ "text", "image", "audio" ], "tools": true, "deployment": "checkpoint", "identityLabel": "checkpoint", "capabilityNote": "Produces text and supports optional thinking and tools. Audio understanding remains experimental. The October 15 article explains an earlier July 24 release; it is not a new October checkpoint.", "accessNote": "The project publishes Apache 2.0 weights and training details. Choose the 8B or 70B download to match your available hardware, and follow the project\u2019s setup guide.", "pricingNote": "No hosted token quote is verified for this exact profile. Local hardware and operating costs are separate.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "nemotron-3-5-bolt-e4b", "name": "Nemotron 3.5 Bolt E4B \u00b7 download", "provider": "NVIDIA", "providerId": "nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-4GB", "sourceUrl": "https://blogs.nvidia.com/blog/nemotron-3-5-super-vl/", "officialUrl": "https://blogs.nvidia.com/blog/nemotron-3-5-super-vl/", "note": "Run a small text assistant locally, including tool calls. A voice assistant still needs separate speech recognition and speech output.", "discovered": "2026-10-16", "released": "2026-10-23", "context": 65536, "maxOutput": null, "modalities": [ "text" ], "tools": true, "announced": "2026-10-15", "verification": "Download verified", "availability": "available", "accessNote": "The downloadable GGUF weights use OpenMDW 1.1. Check the license and your runtime\u2019s memory requirements. Publisher speed and quality claims are not independent measurements here.", "pricingNote": "No hosted rate is verified.", "capabilityNote": "Thinking is enabled by default and can be disabled with enable_thinking. A runtime token budget is not a verified low/medium/high effort menu.", "deployment": "checkpoint", "checkpointRevision": "1db54d5b12533ff1bcaa8e42cf4c667988a1f1be", "contextNote": "The model card documents 65,536 tokens shared by input and output. A 4GB weight target does not guarantee that a full-length request fits in 4GB of total memory.", "apiDocumentationUrls": [ "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-4GB", "https://research.nvidia.com/labs/nemotron/NVIDIA-Nemotron-3.5-Bolt-E4B-4GB/" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "nemotron-3-5-bolt-74b", "name": "Nemotron 3.5 Bolt 74B \u00b7 24GB", "provider": "NVIDIA", "providerId": "nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB", "sourceUrl": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB", "officialUrl": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB", "note": "Run text and visual tasks on your own hardware with NVIDIA\u2019s compressed 74B model. The 24GB package reduces the weight footprint; working memory and a compatible serving setup are still needed.", "discovered": "2026-10-16", "released": "2026-10-15", "context": 262144, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": true, "deployment": "checkpoint", "identityLabel": "checkpoint", "accessNote": "OpenMDW 1.1 weights. Follow the provider\u2019s supported GPU and software recipe; the package size is not a guarantee that the full context fits on every 24GB device.", "contextNote": "The card supports up to 262,144 combined tokens. Prompts, images, video, history and generated text share the space; serving recipes can set smaller limits.", "capabilityNote": "Thinking can be switched on or off. No matched independent comparison of these modes is available here.", "pricingNote": "This is a download, not a verified hosted token offer. Include hardware, memory and operation in your budget.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "lumma-fev-27b", "providerId": "FrontiersMind/Lumma-fev-27B", "identityLabel": "checkpoint", "name": "Lumma-Fev-27B", "provider": "FrontiersMind", "released": "2026-10-14", "discovered": "2026-10-15", "deployment": "checkpoint", "specialized": "decision", "context": 8192, "maxOutput": null, "questionLimit": null, "modalities": [ "text", "image" ], "tools": false, "checkpointRevision": "2c080fb4b44bbad2a140353393d935e8b7af8bcb", "sourceUrl": "https://huggingface.co/FrontiersMind/Lumma-fev-27B", "officialUrl": "https://huggingface.co/blog/FrontiersMind/lumma-fev-27b", "note": "Choose labels or yes/no answers for records in English, Hindi, Tamil and Kannada. This downloadable model returns probabilities and can serve a ticket-routing workflow on your own hardware.", "capabilityNote": "Built from Qwen3.8-27B. The provider trained and tested yes/no and choice questions; ordered score questions are accepted but untested. Image input requires the optional image package. Published comparisons are the developer\u2019s evaluations, not independent evidence.", "contextNote": "Shared information plus one question must fit within 8,192 tokens. Each question reads the shared information separately; answers do not change one another.", "accessNote": "Apache 2.0 weights. The model card recommends about 55 GB of free GPU memory, typically an 80 GB card, or several GPUs. Install the provider\u2019s software and test confidence thresholds on your own examples before automating decisions.", "pricingNote": "This is a local download, with no verified hosted token rate. Include hardware, electricity and maintenance when comparing it with an online API.", "apiDocumentationUrls": [ "https://huggingface.co/FrontiersMind/Lumma-fev-27B/blob/main/LICENSE" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "provider": "Perplexity", "released": null, "discovered": "2026-10-15", "specialized": "decision", "tools": false, "maxOutput": null, "modalities": [ "text", "image" ], "questionLimit": null, "outputType": "Yes/no probabilities, choices and ordered scores", "officialUrl": "https://huggingface.co/perplexity-ai/pplx-decider-v1.1-27b", "slug": "perplexity-decider-v1-1-27b-local", "providerId": "perplexity-ai/pplx-decider-v1.1-27b", "identityLabel": "checkpoint", "name": "Perplexity Decider v1.1 27B \u00b7 local download", "deployment": "checkpoint", "context": 8192, "checkpointRevision": "14c9d90137962d300f3de95633f7b6535e80c5ce", "sourceUrl": "https://huggingface.co/perplexity-ai/pplx-decider-v1.1-27b", "note": "Run the v1.1 decision model on your own GPU to assign labels and probabilities to text or images. This download has different limits from the hosted Gateway offer.", "capabilityNote": "The supplied implementation reads each decision within 8,192 tokens and rejects longer inputs. Images share that space. Its training and reported tests use smaller images than the default serving setting; use the documented setup when reproducing those results.", "contextNote": "The supplied prepare function limits text and all images for one decision to 8,192 tokens. Training and evaluation used images up to 512\u00d7512; the serving default allows up to 2048\u00d72048 and uses more memory.", "accessNote": "Weights are Apache 2.0. The card asks for Python 3.12+, CUDA and roughly 49 GiB for weights plus working memory. Its private-repository wording conflicts with publicly readable files; confirm download access. Keep the supplied attention and calibration settings to preserve the tested behavior.", "pricingNote": "No hosted token price applies to this download. Budget for hardware, memory and operation; the Gateway quote does not measure local running costs.", "apiDocumentationUrls": [ "https://huggingface.co/perplexity-ai/pplx-decider-v1.1-27b/blob/main/LICENSE" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "provider": "Perplexity", "released": null, "discovered": "2026-10-15", "specialized": "decision", "tools": false, "maxOutput": null, "modalities": [ "text", "image" ], "questionLimit": null, "outputType": "Yes/no probabilities, choices and ordered scores", "officialUrl": "https://huggingface.co/perplexity-ai/pplx-decider-v1.1-27b", "slug": "perplexity-decider-v1-1-27b", "providerId": "perplexity/pplx-decider-v1.1-27b", "identityLabel": "AI Gateway route", "name": "Perplexity Decider v1.1 27B \u00b7 Gateway", "context": null, "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "verification": "No longer listed on Gateway", "note": "This older Gateway offer is no longer in the current catalog. For a new project, start with Decider v1.2 and test it on decisions you have already labeled.", "capabilityNote": "The former Gateway offer accepted text and images for classification, routing and scoring. Downloaded v1.1 weights remain a separate way to use the older model.", "contextNote": "Gateway last listed 262,144 context tokens on October 17. That is a historical hosting limit; the downloadable implementation uses 8,192 tokens.", "pricingNote": "The last observed Gateway price was $0.02 per million input tokens with free output on October 17. Current availability and pricing are unconfirmed, so this offer is excluded from the cost calculator.", "billingNote": "Historical Gateway pricing is not a current quote.", "accessNote": "Gateway removed this exact ID from its October 18 catalog. Perplexity\u2019s native API redirects older v1 and v1.1 names to v1.2. Its response echoes the requested name, so the name alone does not prove which version ran.", "availability": "unlisted", "dateNote": "Current access, prices and serving limits are unconfirmed. The details below describe the last listed offer.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "scaledown-classify-1", "providerId": "scaledown/scaledown-classify-1", "identityLabel": "AI Gateway route", "name": "ScaleDown Classify-1", "provider": "ScaleDown", "discovered": "2026-10-14", "verification": "Specialized Gateway service", "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": false, "specialized": "classification", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "officialUrl": "https://docs.scaledown.ai/api-reference/classify-overview", "note": "Assign text to your own labels, such as sales, support or billing, and receive a probability for each label.", "capabilityNote": "The provider supports custom labels with rubrics and a separate fixed-answer Decisions API. This profile covers the Gateway Classify-1 offer; general chat scores do not establish classification accuracy. In the native classification API, the optional explanation is a second call enabled by default, with its input billed at twice the base rate. It explains the chosen label; it is not an independent second opinion. Gateway\u2019s exact handling of that option is unverified.", "documentedRates": { "input": 0.05, "output": 0, "cached": null, "label": "AI Gateway offer \u00b7 USD" }, "pricingNote": "Gateway lists $0.05 per million input tokens and free output. Its zero capacity fields do not establish a usable zero-token limit; request limits, caching and other fees are unverified.", "accessNote": "Test examples with known labels, especially ambiguous cases that should go to a person. A high confidence score alone does not establish that an answer is correct.", "apiDocumentationUrls": [ "https://docs.scaledown.ai/api-reference/classify", "https://docs.scaledown.ai/api-reference/decisions" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "released": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "nemotron-3-5-lightning-free", "providerId": "nvidia/nemotron-3.5-lightning-free", "identityLabel": "AI Gateway route", "name": "Nemotron 3.5 Lightning \u00b7 free Gateway offer", "provider": "NVIDIA", "released": "2026-08-11", "discovered": "2026-10-14", "verification": "Newly covered hosted offer", "context": 1000000, "maxOutput": 65000, "modalities": [ "text" ], "tools": true, "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "officialUrl": "https://developer.nvidia.com/blog/nvidia-nemotron-3-5-lightning-delivers-fast-accurate-specialized-task-execution-for-long-running-agents/", "note": "A newly listed free Gateway offer for NVIDIA\u2019s August model. It supports text assistants, tools and long documents.", "capabilityNote": "The hosted offer lists a 1M context, 65,000 output tokens and structured answers. The 30B downloadable model uses about 3B parameters per token; download hardware limits and licenses are separate.", "documentedRates": { "input": 0, "output": 0, "cached": null, "label": "AI Gateway free offer \u00b7 USD" }, "pricingNote": "Gateway currently lists free input and output. Promotion duration, usage quotas and caching rates are unverified; this is not a promise of permanent free access.", "accessNote": "Gateway states that the provider may retain prompts and responses for model improvement. Consider that data policy before sending private work.", "gatewayEffortOptions": [ "none", "low", "medium", "high" ], "capabilitySourceUrl": "https://ai-gateway.vercel.sh/v1/models", "apiDocumentationUrls": [ "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gemini-4-argon-cyber", "name": "Gemini 4 Argon \u00b7 cyber defense version", "provider": "Google", "providerId": null, "identityLabel": "Restricted model", "released": null, "discovered": "2026-10-14", "verification": "Restricted Fairwind access", "context": null, "maxOutput": null, "modalities": [ "text" ], "tools": null, "sourceUrl": "https://deepmind.google/models/gemini/cyber/", "officialUrl": "https://blog.google/innovation-and-ai/technology/safety-security/gemini-4-model-safety-security/", "note": "Google offers a specialized Argon version to approved cyber defenders for finding vulnerabilities and preparing fixes.", "capabilityNote": "Google describes vulnerability discovery, penetration testing and patching. Those are provider claims; no independent matched result is assigned to this restricted version.", "accessNote": "Access is limited to trusted defenders in the Fairwind Program. A public API ID, ordinary account access and this version\u2019s token limits are unverified. For new paid Gemini API projects from October 26, Generate Content requests are stored by default. Existing projects keep their setting; opt out with store=false or the logging dashboard.", "pricingNote": "Separate prices have not been confirmed. Public Argon prices and its Arena result are not assigned to this version.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "provider": "Nace.AI", "discovered": "2026-10-13", "context": 131072, "maxOutput": null, "modalities": [ "text" ], "tools": false, "specialized": "decision", "questionLimit": null, "outputType": "Yes/no probabilities, multiple-choice probabilities and ordered rating scores. No generated explanation.", "released": null, "slug": "drex-v2-0-local", "providerId": "nace-ai/drex-v2.0", "identityLabel": "checkpoint", "name": "Drex v2.0 \u00b7 local checkpoint", "deployment": "checkpoint", "verification": "Local decision checkpoint", "sourceUrl": "https://huggingface.co/nace-ai/drex-v2.0", "officialUrl": "https://github.com/nace-ai/drex-decision-models/tree/main/models/drex-v2.0", "checkpointRevision": "a21d4daec2d114bf40058d096373329549dd8b3b", "note": "Run typed decisions on your own hardware. This 35B model uses about 3B parameters per token; it returns option probabilities rather than prose.", "capabilityNote": "Built from Qwen3.6-35B-A3B with a decision head and text input. The provider\u2019s Python, llama.cpp and Ollama runners support it. Publisher performance claims remain separate from independent results.", "contextNote": "The documented runner default is 16,384 tokens, with an explicitly enabled maximum of 131,072. In llama-server, context is split across slots: the example -np 2 / -c 32768 gives 16,384 per slot. Native API row and question limits do not establish local runner limits.", "accessNote": "Weights use the modified Nace.AI Open RAIL-M license, while runtime code is Apache 2.0. The weight license restricts competing products and generally requires a separate license for non-research/non-personal use above $1 million prior-year gross revenue or total funding. Read the exact license and its other use restrictions before deployment.", "pricingNote": "The publisher lists about 70 GB BF16 or 37 GB Q8_0 weights before runtime and context overhead. No hosted token price or measured running cost is assigned.", "apiDocumentationUrls": [ "https://huggingface.co/nace-ai/drex-v2.0/blob/a21d4daec2d114bf40058d096373329549dd8b3b/LICENSE", "https://github.com/nace-ai/drex-decision-models/blob/main/docs/context-length.md" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "provider": "Nace.AI", "discovered": "2026-10-13", "context": 131072, "maxOutput": null, "modalities": [ "text" ], "tools": false, "specialized": "decision", "questionLimit": 512, "outputType": "Yes/no probabilities, multiple-choice probabilities and ordered rating scores. No generated explanation.", "released": "2026-10-06", "slug": "drex-v2-0", "providerId": "drex-v2.0", "name": "Drex v2.0 \u00b7 native API", "verification": "Native API contract verified", "sourceUrl": "https://console.nace.ai/docs/drex/reference/models", "officialUrl": "https://console.nace.ai/docs/drex/reference/models", "apiDocumentationUrls": [ "https://console.nace.ai/docs/drex/reference/pricing", "https://console.nace.ai/docs/drex/reference/limits" ], "documentedRates": { "input": 0.06, "output": 0, "cached": null, "label": "Native Drex API \u00b7 USD" }, "note": "Choose from allowed answers without generating a paragraph: route a ticket, classify a document or return a probability for each option.", "capabilityNote": "Send text state and up to 512 typed questions to /v1/systemone. Drex v2.0 returns probabilities, not a written explanation. Validate those probabilities on examples from your own workflow.", "contextNote": "Drex v2.0 accepts up to 131,072 tokens of shared information, and up to 262,144 including the longest question. It allows up to 512 questions and 4,194,304 bytes for the serialized request. Requests above a limit are refused; v2.0 no longer hands long requests to v1.5.", "accessNote": "The provider dates v2.0 October 6 and moved drex-latest and omitted-model requests to it on October 12. Use the exact ID drex-v2.0 to select this version. The legacy nacedm-latest alias selects the current default. Other unknown model names now return HTTP 422 instead of silently selecting a model.", "pricingNote": "Native v2.0 costs $0.06 per million billed input tokens, with no output, request or question fee. Only successful evaluations are charged. Use the returned usage totals to compare bills.", "billingNote": "Native v2.0 charges $0.06 per million billed input tokens. Requests over its input limits are refused instead of being answered by v1.5.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "muse-wonder-full", "name": "Muse Wonder Full", "provider": "Meta", "providerId": null, "identityLabel": "Subscription model", "released": "2026-10-12", "discovered": "2026-10-13", "verification": "Provider verified \u00b7 subscription access", "context": 1000000, "maxOutput": null, "modalities": [ "text" ], "tools": null, "sourceUrl": "https://dev.meta.ai/models/muse-wonder", "officialUrl": "https://dev.meta.ai/docs/models", "note": "Meta\u2019s higher-capability Wonder variant for work in Muse Code. It requires a Power Usage or Maximum Usage subscription.", "capabilityNote": "Meta positions Full for demanding coding and longer tasks. Its published comparisons do not establish an independently verified score in our matched datasets.", "contextNote": "The product page advertises 1M context. A precise token limit and maximum output for this subscription variant are not verified.", "accessNote": "Full is not available through the pay-as-you-go Model API. Choosing max thinking on ordinary Wonder does not unlock it. Muse Code subscriptions default to Contributor mode, allowing submitted content to train future models; Standard mode is available in settings. Meta currently excludes South Korea from Wonder Full access.", "pricingNote": "A subscription is required. No per-token price is assigned, and ordinary Wonder\u2019s API prices and benchmark score do not apply.", "deployment": "subscription", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "mellum-2-1-thinking", "provider": "JetBrains", "providerId": "JetBrains/Mellum2.1-12B-A2.5B-Thinking", "identityLabel": "checkpoint", "name": "Mellum2.1 Thinking", "deployment": "checkpoint", "released": "2026-10-08", "discovered": "2026-10-12", "verification": "Publisher checkpoint verified", "context": 131072, "maxOutput": null, "modalities": [ "text" ], "tools": true, "checkpointRevision": "92ddae9fc7665e9f801d141d2e5a6b2caf2460c4", "sourceUrl": "https://huggingface.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking", "officialUrl": "https://blog.jetbrains.com/ai/2026/10/mellum2-1-gets-to-work-a-fast-open-model-for-coding-agents/", "note": "A coding model you can run on your own hardware. Use it to investigate a failing test, draft a fix and check the result while keeping repository access under your control.", "capabilityNote": "JetBrains describes a 12B model with 2.5B parameters active at each step. It supports reasoning and tool calls; your application supplies the shell, files and permissions. Its published benchmark results are vendor measurements, so they are not mixed into our independent rankings.", "contextNote": "The model card specifies 131,072 tokens of context. Its example output budget is a request setting, not a verified maximum output limit. Allow room for the prompt, tool results and generated answer.", "accessNote": "The publisher labels the checkpoint Apache 2.0. The card documents vLLM setup with reasoning and tool parsers. Official GGUF files are now available, although the announcement still says they are coming soon; choose a format supported by your local runner.", "pricingNote": "No verified hosted token price or measured running cost is assigned. Local use still requires hardware, memory and setup. Test changes in a disposable copy of your project before granting access to important files.", "apiDocumentationUrls": [ "https://huggingface.co/JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "cloudflare-clef-omni", "providerId": "@cf/cloudflare/clef-omni", "name": "Cloudflare Clef Omni", "provider": "Cloudflare", "discovered": "2026-10-11", "released": null, "context": 64000, "maxOutput": null, "modalities": [ "text", "image", "audio", "video" ], "tools": false, "specialized": "decision", "questionLimit": 64, "outputType": "Yes/no probabilities, choice probabilities and ordered rubric scores; no generated explanation.", "sourceUrl": "https://developers.cloudflare.com/workers-ai/models/clef-omni/", "officialUrl": "https://huggingface.co/Cloudflare/clef-omni", "verification": "Hosted multimodal decision model", "documentedRates": { "input": 0.15, "output": 0, "cached": null, "label": "Workers AI documented rates \u00b7 USD" }, "note": "Classify or route records that include images, audio or video. Clef Omni returns allowed answers with probabilities instead of prose, using a 30B mixture-of-experts backbone with 3B active parameters.", "capabilityNote": "The hosted schema explicitly accepts up to 64 typed questions, four embedded images, four audio clips and two videos. Audio clips can be 300 seconds / 8 MiB each; videos 60 seconds / 16 MiB each, sampled at 2 frames / second. Audio and video total at most 16 MiB decoded. Remote media URLs are not accepted.", "contextNote": "The hosted context is exactly 64,000 tokens, shared by media, questions and text. If media and questions exceed it, the request fails. Otherwise text state is truncated to fit. This is separate from Clef\u2019s 65,536 and Flash\u2019s 24,576 limits.", "pricingNote": "Input is $0.15 per million billed tokens; output tokens are free. Images add up to 1,024 tokens each, audio about 780 tokens / minute, and video up to about 15,400 frame tokens/minute plus audio. These are provider tokenization estimates, not measured task costs. Cache rates and other charges remain unverified.", "billingNote": "Workers AI explicitly documents free output. Include media and question tokens in measured billed input; video token use depends on duration and resolution.", "accessNote": "Hosted through a Cloudflare account. The separate checkpoint card declares Apache 2.0 and Qwen3-Omni-30B-A3B-Instruct as its parent, pinned at 0db1cd2607d76a7bdb2a382f659e7b313079f84b. Repository creation October 9 is an artifact date; a provider launch date has not been verified. Card benchmark tables are vendor runs and do not become independent general-model scores.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "provider": "Nace.AI", "discovered": "2026-10-11", "context": 32768, "maxOutput": null, "modalities": [ "text" ], "tools": false, "specialized": "decision", "questionLimit": null, "outputType": "Yes/no probabilities, multiple-choice probabilities and ordered rating scores. No generated explanation.", "released": null, "slug": "drex-dlm", "providerId": "nace-ai/drex-dlm", "identityLabel": "checkpoint", "name": "Drex DLM \u00b7 local checkpoint", "deployment": "checkpoint", "verification": "Local diffusion decision checkpoint", "sourceUrl": "https://huggingface.co/nace-ai/drex-dlm", "officialUrl": "https://github.com/nace-ai/drex-decision-models/tree/5c3d2220713c732a6820c4cdda51c7c38db991f5/models/drex-dlm", "checkpointRevision": "8827b917a7e1cb037ce4fcef84f814a16158d9ad", "note": "A local decision checkpoint built from Efficient-DLM-8B. It reads shared state and returns probabilities for typed questions without generating prose.", "capabilityNote": "The shared state uses bidirectional attention, with separate causal question branches and a pointer head. Requests inside the packed-token budget use one forward pass; larger ones split into question rows. This is a different checkpoint from Drex v1.5.", "contextNote": "The exact card documents 32,768 maximum tokens, enabled explicitly; runners default to 16,384. Align the context, batch and microbatch caps. The provider has checked Q8_0 at 16,384; its full-window equivalence is not established by that test.", "accessNote": "Weights are CC BY-NC 4.0, which does not grant ordinary commercial use. Runtime code is separately Apache 2.0. The provider reports tests on Apple M5 Max/M5 Pro and an H100; CPU-only inference is untested. These are vendor compatibility checks, not independently measured performance.", "pricingNote": "Native v1.5\u2019s API price does not apply to this local diffusion checkpoint. About 16 GB BF16 weights are documented before overhead; hosted pricing and measured operating cost remain unknown. Repository creation October1 and later modification are artifact dates, not a verified launch date.", "apiDocumentationUrls": [ "https://github.com/nace-ai/drex-decision-models/blob/5c3d2220713c732a6820c4cdda51c7c38db991f5/models/drex-dlm/LICENSE.md" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "provider": "Nace.AI", "discovered": "2026-10-11", "context": 131072, "maxOutput": null, "modalities": [ "text" ], "tools": false, "specialized": "decision", "questionLimit": null, "outputType": "Yes/no probabilities, multiple-choice probabilities and ordered rating scores. No generated explanation.", "released": null, "slug": "drex-v1-5-local", "providerId": "nace-ai/drex-v1.5", "identityLabel": "checkpoint", "name": "Drex v1.5 \u00b7 local checkpoint", "deployment": "checkpoint", "verification": "Local decision checkpoint", "sourceUrl": "https://huggingface.co/nace-ai/drex-v1.5", "officialUrl": "https://github.com/nace-ai/drex-decision-models/tree/5c3d2220713c732a6820c4cdda51c7c38db991f5/models/drex-v1.5", "checkpointRevision": "328ee227c0425669d9f37fa82d400aa52c22139c", "note": "Run the same typed decision format on your own hardware. This is the local 9B checkpoint, separate from the billed native API.", "capabilityNote": "The model uses MiMo-V2.6-Distill-Qwen-9B and a pointer head. Python, the provider\u2019s llama.cpp fork and its Ollama fork support /v1/systemone. Each question is scored in one pass; no text is generated.", "contextNote": "The documented runner default is 16,384 tokens, with an explicitly enabled maximum of 131,072. In llama-server, context is split across slots: the example -np 2 / -c 32768 gives 16,384 per slot. Native API row and question limits do not establish local runner limits.", "accessNote": "Weights use the modified Nace.AI Open RAIL-M license, while runtime code is Apache 2.0. The weight license restricts competing products and generally requires a separate license for non-research/non-personal use above $1 million prior-year gross revenue or total funding. Read the exact license and its other use restrictions before deployment.", "pricingNote": "No hosted token rate or measured operating cost is assigned. The provider describes about 18 GB BF16 weights or 9.5 GB Q8_0, before runtime and context overhead. These are artifact estimates, not measured peak memory or latency.", "apiDocumentationUrls": [ "https://huggingface.co/nace-ai/drex-v1.5/blob/328ee227c0425669d9f37fa82d400aa52c22139c/LICENSE", "https://github.com/nace-ai/drex-decision-models/blob/5c3d2220713c732a6820c4cdda51c7c38db991f5/docs/context-length.md" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "provider": "Nace.AI", "discovered": "2026-10-11", "context": 131072, "maxOutput": null, "modalities": [ "text" ], "tools": false, "specialized": "decision", "questionLimit": 512, "outputType": "Yes/no probabilities, multiple-choice probabilities and ordered rating scores. No generated explanation.", "released": "2026-09-28", "slug": "drex-v1-5", "providerId": "drex-v1.5", "name": "Drex v1.5 \u00b7 native API", "verification": "Native API contract verified", "sourceUrl": "https://console.nace.ai/docs/drex/reference/models", "officialUrl": "https://github.com/nace-ai/drex-decision-models/tree/5c3d2220713c732a6820c4cdda51c7c38db991f5/models/drex-v1.5", "apiDocumentationUrls": [ "https://console.nace.ai/docs/drex/reference/pricing", "https://console.nace.ai/docs/drex/reference/limits" ], "documentedRates": { "input": 0.05, "output": 0, "cached": null, "label": "Native Drex API \u00b7 USD" }, "note": "Turn a document into allowed answers: route a ticket, classify a record or score a condition without generating a paragraph. This profile covers the native Nace API.", "capabilityNote": "Send model drex-v1.5 to POST /v1/systemone at console.nace.ai. The provider documents text state and yes/no, choice and rating questions. It returns probabilities; validate calibration on your own records.", "contextNote": "State alone can contain 131,072 tokens; state plus the longest question can contain 139,264. Other questions do not add to that row limit. A request allows up to 512 questions and 4,194,304 serialized UTF-8 bytes for state plus questions. Check all boundaries.", "accessNote": "Send drex-v1.5 to keep this version. On October 12, the default and drex-latest moved to v2.0. Retired drex-v1.1 still falls back to v1.5. Current v2.0 refuses requests above its own limits instead of handing them to v1.5.", "pricingNote": "Native v1.5 is $0.05 per million billed input tokens, with no output-token or per-request/per-question charge. Only successful HTTP 200 evaluations are debited. Use usage.input_tokens from actual responses; the OpenRouter page advertises a different $0.04 offer, but its exact route is absent from both reviewed catalogs, so access and billing there remain unverified.", "billingNote": "Native output tokens are free; no request or question fee. The estimate uses billed usage, before account credits. A retired ID is billed at the version that actually answers.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "firerouter-opus", "providerId": "firerouter/opus", "name": "FireRouter with Opus", "provider": "Fireworks", "released": "2026-09-28", "announced": "2026-09-28", "discovered": "2026-10-10", "verification": "Custom route removed", "context": null, "maxOutput": null, "modalities": [ "text", "image" ], "tools": true, "sourceUrl": "https://fireworks.ai/models/fireworks/firerouter-opus", "officialUrl": "https://fireworks.ai/blog/introducing-firerouter-with-opus", "apiDocumentationUrls": [ "https://docs.fireworks.ai/nexus/firerouter", "https://docs.fireworks.ai/nexus/firerouter/setup" ], "note": "This former Opus router is no longer available. Fireworks now documents Nexus Auto for model selection, or a specific model ID to keep every turn on one model.", "capabilityNote": "Fireworks says firerouter, firerouter/opus and auto-instant are no longer available. The old custom route does not have a current guaranteed model mix.", "accessNote": "Migrate to auto for routing, or pin a supported model ID. Recheck credentials, application behavior and billing before switching.", "contextNote": "No current capacity is established for this removed route.", "pricingNote": "Historical route prices are not a current offer. Nexus Auto bills according to the model that serves each turn.", "availability": "unlisted", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-10", "maxOutput": null, "tools": null, "context": null, "modalities": [ "text" ], "slug": "falcon-emirati-7b", "providerId": null, "name": "Falcon Emirati 7B", "provider": "TII / Falcon", "announced": "2026-10-06", "released": null, "verification": "Announced dialect specialist", "specialized": "dialect", "sourceUrl": "https://huggingface.co/blog/tiiuae/falcon-emirati", "officialUrl": "https://chat.falconllm.tii.ae/?model=Falcon-Emirati-7B", "note": "A text model adapted for Emirati Arabic, including local vocabulary, idioms and cultural context. It can be useful when a general Arabic assistant answers in the wrong dialect.", "capabilityNote": "The provider\u2019s announcement names Falcon-Emirati-7B and links a public chat. The native API ID, downloadable checkpoint and tool contract remain unverified; the publisher model search returned no exact repository.", "contextNote": "Do not borrow the 128K/256K context limits of the wider Falcon-H1-Arabic family. A limit for this exact dialect model has not been verified.", "accessNote": "The provider reports native-speaker reviews, an Emirati-language benchmark and AI judging. Those are provider evidence. Test naturalness, dialect consistency and correct answers with your own native speakers before selecting it.", "pricingNote": "Exact-model API pricing and a checkpoint license are unverified. Parent-model license, effort settings and benchmark scores are not assigned to this release.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-10", "maxOutput": null, "tools": false, "context": null, "modalities": [ "text", "image" ], "slug": "lightonocr-3-4b", "providerId": "lightonai/LightOnOCR-3-4B", "identityLabel": "checkpoint", "name": "LightOnOCR-3 4B", "provider": "LightOn", "announced": "2026-10-08", "released": null, "verification": "Document extraction checkpoint", "specialized": "document-extraction", "deployment": "checkpoint", "sourceUrl": "https://huggingface.co/lightonai/LightOnOCR-3-4B", "officialUrl": "https://huggingface.co/blog/lightonai/lightonocr-3", "checkpointRevision": "b71010f095dc3735de7ee6b9969bff8f8f50b39a", "note": "Read page images into text, tables and labeled layout regions. This is the larger checkpoint; test it against the smaller versions on your documents.", "capabilityNote": "Use an empty prompt for transcription or the grounding prompt for labeled bounding boxes, image descriptions and chart data. Other instructions are outside the documented training distribution; it is not a general chat or tool assistant.", "contextNote": "The card\u2019s sample output budget is not a published maximum. Image resolution, page complexity and output length change memory use and throughput; the supported document/request cap is unverified.", "accessNote": "The card declares Apache 2.0. The 0.8B and 4B variants use Qwen3.5 vision-language architecture, while 1B keeps the previous architecture. Provider OCR tables and normalization choices are not imported as independent general-purpose scores.", "pricingNote": "Hosted prices and measured costs for your documents are unknown. Compare extraction errors, missing chart values, page latency and total operating cost using the same rendered pages.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-10", "maxOutput": null, "tools": false, "context": null, "modalities": [ "text", "image" ], "slug": "lightonocr-3-1b", "providerId": "lightonai/LightOnOCR-3-1B", "identityLabel": "checkpoint", "name": "LightOnOCR-3 1B", "provider": "LightOn", "announced": "2026-10-08", "released": null, "verification": "Document extraction checkpoint", "specialized": "document-extraction", "deployment": "checkpoint", "sourceUrl": "https://huggingface.co/lightonai/LightOnOCR-3-1B", "officialUrl": "https://huggingface.co/blog/lightonai/lightonocr-3", "checkpointRevision": "23d544085c886bbdace647bcfab91203299417e2", "note": "Read page images into text, tables and labeled layout regions. This variant keeps the earlier LightOnOCR-2 architecture.", "capabilityNote": "Use an empty prompt for transcription or the grounding prompt for labeled bounding boxes, image descriptions and chart data. Other instructions are outside the documented training distribution; it is not a general chat or tool assistant.", "contextNote": "The card\u2019s sample output budget is not a published maximum. Image resolution, page complexity and output length change memory use and throughput; the supported document/request cap is unverified.", "accessNote": "The card declares Apache 2.0. The 0.8B and 4B variants use Qwen3.5 vision-language architecture, while 1B keeps the previous architecture. Provider OCR tables and normalization choices are not imported as independent general-purpose scores.", "pricingNote": "Hosted prices and measured costs for your documents are unknown. Compare extraction errors, missing chart values, page latency and total operating cost using the same rendered pages.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-10", "maxOutput": null, "tools": false, "context": null, "modalities": [ "text", "image" ], "slug": "lightonocr-3-0-8b", "providerId": "lightonai/LightOnOCR-3-0.8B", "identityLabel": "checkpoint", "name": "LightOnOCR-3 0.8B", "provider": "LightOn", "announced": "2026-10-08", "released": null, "verification": "Document extraction checkpoint", "specialized": "document-extraction", "deployment": "checkpoint", "sourceUrl": "https://huggingface.co/lightonai/LightOnOCR-3-0.8B", "officialUrl": "https://huggingface.co/blog/lightonai/lightonocr-3", "checkpointRevision": "36d8636925327c671bd8e479c517a004f7566bd9", "note": "Read page images into text, tables and labeled layout regions. This is the smallest checkpoint in the family.", "capabilityNote": "Use an empty prompt for transcription or the grounding prompt for labeled bounding boxes, image descriptions and chart data. Other instructions are outside the documented training distribution; it is not a general chat or tool assistant.", "contextNote": "The card\u2019s sample output budget is not a published maximum. Image resolution, page complexity and output length change memory use and throughput; the supported document/request cap is unverified.", "accessNote": "The card declares Apache 2.0. The 0.8B and 4B variants use Qwen3.5 vision-language architecture, while 1B keeps the previous architecture. Provider OCR tables and normalization choices are not imported as independent general-purpose scores.", "pricingNote": "Hosted prices and measured costs for your documents are unknown. Compare extraction errors, missing chart values, page latency and total operating cost using the same rendered pages.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-10", "maxOutput": null, "tools": false, "context": null, "modalities": [ "text" ], "slug": "darwin-27b-ztc-v2", "providerId": "FINAL-Bench/Darwin-27B-ZTC-v2", "identityLabel": "checkpoint", "name": "Darwin 27B ZTC v2", "provider": "FINAL-Bench", "released": null, "verification": "Open-weight decision checkpoint", "specialized": "decision", "deployment": "checkpoint", "questionLimit": null, "sourceUrl": "https://huggingface.co/FINAL-Bench/Darwin-27B-ZTC-v2", "officialUrl": "https://huggingface.co/FINAL-Bench/Darwin-27B-ZTC-v2", "checkpointRevision": "ed78457802e7825f3f6c7fa7492d67feb9124ae2", "dateNote": "The repository was created October 8; a separate provider launch date is unverified.", "note": "The second Darwin decision checkpoint adds training for control and workflow choices. Keep it separate from v1 when testing accuracy or calibration.", "capabilityNote": "The supplied interface supports yes/no, choice and ordered-score questions without generating answer tokens. It uses a text-only readout head on the Darwin backbone; use the matching inference code and checkpoint.", "contextNote": "The configuration lists 262,144 internal positions, but a validated maximum for the decision interface is unverified. The v2 benchmark used an explicit 32,768 limit for one task; that is a test setting, not a universal endpoint allowance.", "accessNote": "The card declares Apache 2.0; bundled autojev code carries its MIT notice. The provider describes about 54 GB of BF16 weights, before runtime memory. v1\u2019s card corrects mixed-run numbers. v2 discloses training on other splits of tasks that also appear in S1MB; we have not audited that leaderboard or imported its scores.", "pricingNote": "No hosted per-token price is verified for this exact checkpoint. Local GPU, memory and operating costs need your own measurement.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-10", "maxOutput": null, "tools": false, "context": null, "modalities": [ "text" ], "slug": "darwin-27b-ztc", "providerId": "FINAL-Bench/Darwin-27B-ZTC", "identityLabel": "checkpoint", "name": "Darwin 27B ZTC", "provider": "FINAL-Bench", "released": null, "verification": "Open-weight decision checkpoint", "specialized": "decision", "deployment": "checkpoint", "questionLimit": null, "sourceUrl": "https://huggingface.co/FINAL-Bench/Darwin-27B-ZTC", "officialUrl": "https://huggingface.co/FINAL-Bench/Darwin-27B-ZTC", "checkpointRevision": "2b4bb9e22687b89e3c5703696b575ea2d2774da8", "dateNote": "The repository was created October 8; a separate provider launch date is unverified.", "note": "A local text decision model that scores fixed choices in one forward pass per question.", "capabilityNote": "The supplied interface supports yes/no, choice and ordered-score questions without generating answer tokens. It uses a text-only readout head on the Darwin backbone; use the matching inference code and checkpoint.", "contextNote": "The configuration lists 262,144 internal positions, but a validated maximum for the decision interface is unverified. The v2 benchmark used an explicit 32,768 limit for one task; that is a test setting, not a universal endpoint allowance.", "accessNote": "The card declares Apache 2.0; bundled autojev code carries its MIT notice. The provider describes about 54 GB of BF16 weights, before runtime memory. v1\u2019s card corrects mixed-run numbers. v2 discloses training on other splits of tasks that also appear in S1MB; we have not audited that leaderboard or imported its scores.", "pricingNote": "No hosted per-token price is verified for this exact checkpoint. Local GPU, memory and operating costs need your own measurement.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-10", "maxOutput": null, "tools": false, "context": 32768, "modalities": [ "text" ], "slug": "microsoft-decision-1", "providerId": "microsoft/microsoft-decision-1", "identityLabel": "AI Gateway route", "name": "Microsoft Decision-1", "provider": "Microsoft", "released": "2026-10-09", "verification": "Hosted decision model \u00b7 Preview", "specialized": "decision", "questionLimit": null, "sourceUrl": "https://ai.azure.com/catalog/models/Microsoft-Decision-1", "officialUrl": "https://commandline.microsoft.com/microsoft-decision-1-model-foundry/", "capabilitySourceUrl": "https://vercel.com/changelog/microsoft-decision-1-now-available-on-ai-gateway", "documentedRates": { "input": 0.042, "output": 0, "cached": null, "label": "Microsoft base rate / matching Gateway quote \u00b7 USD" }, "note": "Classify a ticket, choose a workflow step or grade an answer against your rubric. Microsoft\u2019s new text-only decision model returns probabilities over the options you supply, rather than writing a response.", "capabilityNote": "Foundry lists Microsoft-Decision-1 version 1 as Preview. The confirmed Gateway route supports yes/no, choice and ordered scoring, with multiple questions sharing input. It needs a Gateway key without a separate Azure deployment.", "contextNote": "Foundry and Gateway agree on 32,768 input tokens. A maximum number of questions and a cache-read price remain unverified. It does not generate explanations or accept images.", "outputType": "Probabilities for supplied yes/no, choice and rating options; no generated explanations", "pricingNote": "Microsoft publishes $0.042 per million input tokens and free output; Gateway matches this quote. Verify how the selected endpoint bills shared input and questions. We have not measured a full invoice or successful-task cost.", "billingNote": "Output tokens are free. Shared input and multiple questions are supported; exact billed usage still needs verification on your endpoint.", "accessNote": "Microsoft names Qwen3.5-9B as the base and publishes its own comparisons. Parent-model scores and vendor speed claims are not independent results for Decision-1. OpenRouter is linked by the announcement but the route is absent from the current general catalog; Foundry and Gateway availability are separately verified.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-09", "maxOutput": null, "tools": null, "slug": "blama-8b", "providerId": "allenai/Llama-3-Blama-8B", "identityLabel": "checkpoint", "name": "Blama 8B", "provider": "Ai2", "announced": "2026-10-07", "released": null, "verification": "Byte-level research checkpoint", "context": null, "modalities": [ "text" ], "sourceUrl": "https://huggingface.co/allenai/Llama-3-Blama-8B", "officialUrl": "https://allenai.org/blog/bolmo-nature", "note": "A research model that reads and writes bytes instead of ordinary subword tokens. Ai2\u2019s October 7 announcement highlights this new checkpoint; repository creation and modification dates are in August, so announcement and artifact dates stay separate.", "capabilityNote": "This is a base autoregressive text checkpoint, rather than a documented tool-using chat assistant. Its card refers to custom Transformers code and counts max_new_tokens in bytes.", "contextNote": "The configuration\u2019s max_position_embeddings is 8,192. That internal position allowance is not treated as a standard token-context limit. Byte budgets are not directly comparable to the token budgets of hosted chat models.", "accessNote": "The card declares Llama 3 Community License. Blama\u2019s card also uses the older Llama-B name and links; the current repository ID is kept exact. Stage 1 artifacts are training starting points, not extra completed base-model releases. Parent Qwen/Llama scores are not transferred.", "pricingNote": "Hosted prices, output caps and reasoning controls for this exact checkpoint are unverified. Downloadable weights do not make inference free.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-09", "maxOutput": null, "tools": null, "slug": "bwen-8b", "providerId": "allenai/Bwen-8B", "identityLabel": "checkpoint", "name": "Bwen 8B", "provider": "Ai2", "announced": "2026-10-07", "released": null, "verification": "Byte-level research checkpoint", "context": null, "modalities": [ "text" ], "sourceUrl": "https://huggingface.co/allenai/Bwen-8B", "officialUrl": "https://allenai.org/blog/bolmo-nature", "note": "A research model that reads and writes bytes instead of ordinary subword tokens. Ai2\u2019s October 7 announcement highlights this new checkpoint; repository creation and modification dates are in August, so announcement and artifact dates stay separate.", "capabilityNote": "This is a base autoregressive text checkpoint, rather than a documented tool-using chat assistant. Its card refers to custom Transformers code and counts max_new_tokens in bytes.", "contextNote": "The configuration\u2019s max_position_embeddings is 32,768. That internal position allowance is not treated as a standard token-context limit. Byte budgets are not directly comparable to the token budgets of hosted chat models.", "accessNote": "The card declares Apache 2.0. Blama\u2019s card also uses the older Llama-B name and links; the current repository ID is kept exact. Stage 1 artifacts are training starting points, not extra completed base-model releases. Parent Qwen/Llama scores are not transferred.", "pricingNote": "Hosted prices, output caps and reasoning controls for this exact checkpoint are unverified. Downloadable weights do not make inference free.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-09", "maxOutput": null, "tools": false, "slug": "liquid-d1-omni-600m", "providerId": "LiquidAI/d1-omni-600M", "identityLabel": "checkpoint", "name": "Liquid d1 Omni 600M", "provider": "Liquid AI", "released": "2026-10-07", "verification": "Open-weight decision checkpoint", "specialized": "decision", "deployment": "checkpoint", "context": 16384, "questionLimit": null, "modalities": [ "text", "image", "audio" ], "sourceUrl": "https://huggingface.co/LiquidAI/d1-omni-600M", "officialUrl": "https://www.liquid.ai/blog/d1-open", "outputType": "Choice, Score and Noul probabilities without generated output tokens.", "note": "An experimental local decision checkpoint for text with images or audio. It can route spoken requests without generating a transcript or free-form answer.", "capabilityNote": "Use the checkpoint\u2019s supplied decision interface and matching preprocessing. Generic Hugging Face chat examples do not establish a supported generative chat contract.", "contextNote": "The request limit is 16,384 combined text/image/audio positions. With images, state and question text are cut to 896 tokens. Audio clips are cut at 30 seconds, and training covers English speaker\u2013assistant requests.", "accessNote": "Weights use the LFM Open License v1.0, with a commercial-use revenue limitation; review the exact checkpoint license before deployment. Provider benchmark and device timings are not independent measurements. Hosted token prices are unknown; local inference has hardware and operating costs.", "pricingNote": "No public hosted per-token quote is verified for this exact checkpoint. It is not the hosted d1 API and does not inherit its $0.04 rate.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-09", "maxOutput": null, "tools": false, "slug": "liquid-d1-3b", "providerId": "LiquidAI/d1-3B", "identityLabel": "checkpoint", "name": "Liquid d1 3B", "provider": "Liquid AI", "released": "2026-10-07", "verification": "Open-weight decision checkpoint", "specialized": "decision", "deployment": "checkpoint", "context": 32768, "questionLimit": null, "modalities": [ "text", "image" ], "sourceUrl": "https://huggingface.co/LiquidAI/d1-3B", "officialUrl": "https://www.liquid.ai/blog/d1-open", "outputType": "Choice, Score and Noul probabilities without generated output tokens.", "note": "A local decision checkpoint for text and images. It returns fixed answers in one pass rather than writing a response.", "capabilityNote": "Use the checkpoint\u2019s supplied decision interface and matching preprocessing. Generic Hugging Face chat examples do not establish a supported generative chat contract.", "contextNote": "The model card documents 32,768 context tokens for d1-3B.", "accessNote": "Weights use the LFM Open License v1.0, with a commercial-use revenue limitation; review the exact checkpoint license before deployment. Provider benchmark and device timings are not independent measurements. Hosted token prices are unknown; local inference has hardware and operating costs.", "pricingNote": "No public hosted per-token quote is verified for this exact checkpoint. It is not the hosted d1 API and does not inherit its $0.04 rate.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-09", "maxOutput": null, "tools": false, "slug": "liquid-d1", "providerId": "d1", "name": "Liquid d1", "provider": "Liquid AI", "released": "2026-10-05", "verification": "Hosted decision model", "specialized": "decision", "context": null, "questionLimit": null, "modalities": [ "text", "image" ], "sourceUrl": "https://www.liquid.ai/blog/d1-decision-model", "officialUrl": "https://docs.liquid.ai/lfm/models/decision-models", "outputType": "Choice distributions, ordered Score ratings and yes/no Noul probabilities; no free-form text.", "documentedRates": { "input": null, "output": 0, "cached": null, "label": "Earlier native offer \u00b7 current rate unverified" }, "note": "The earlier Liquid decision model remains listed by hosting services. For a new direct Liquid integration, check the newer default d1.1.", "capabilityNote": "Current native documentation lists d1.1 and d1-ultra. Earlier d1 remains on OpenRouter and Gateway; continued native access under the old ID is not established.", "contextNote": "Gateway lists liquid/d1 with 65,536 context and its October 9 announcement explicitly supports images. This is newer than the text-only routed-access note in Liquid\u2019s October 5 launch post. The native maximum context remains unverified; max_tokens=0 describes typed answers rather than zero input capacity.", "pricingNote": "The public pricing page labels d1 at $0.02 per million input tokens, but the native guide now lists d1.1. Confirm the exact service and model before using that quote; the old native profile is excluded from current cost estimates.", "billingNote": "Each question repeats its billable prompt and images. Zero generated tokens means no output-token charge, not free requests.", "accessNote": "Provider comparisons use one run per model on six applications, default chat-model effort, list prices and no cache discounts. They are vendor evidence, not a matched independent quality or latency ranking. The October 9 Gateway announcement confirms typed image questions; check preprocessing and billed image usage on your selected endpoint.", "capabilitySourceUrl": "https://vercel.com/changelog/liquid-ai-d1-is-available-on-ai-gateway", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-11-03", "discovered": "2026-10-09", "maxOutput": null, "tools": null, "slug": "callstack-apex", "providerId": "apex-20261001", "identityLabel": "pinned native model", "name": "Callstack Apex", "provider": "Callstack", "released": "2026-10-01", "verification": "Specialized hosted model", "context": 262144, "modalities": [ "text" ], "sourceUrl": "https://apex.callstack.com/models", "officialUrl": "https://www.callstack.com/blog/apex-agetic-coding-for-mobile-and-web-spcialied-in-react", "capabilitySourceUrl": "https://ai-gateway.vercel.sh/v1/models", "documentedRates": { "input": 0.5, "output": 3, "cached": 0.2, "label": "AI Gateway offer \u00b7 USD" }, "note": "A model trained for React Native and React engineering, with Next.js use described by the provider. The native pinned version is October 1; Gateway availability is a separate October 8 offer. Comparable independent coding results are pending.", "capabilityNote": "The native contract lists text input and a 262,144-token context. Use apex-20261001 for a stable snapshot; apex is a moving alias. Gateway separately lists callstack/apex with text/image input, tools and 131,072 maximum output; those are hosted-route declarations, not verified native capabilities.", "gatewayEffortOptions": [ "none", "low", "medium", "xhigh" ], "contextNote": "The native maximum output and effort default remain unverified. Native text-only documentation differs from Gateway\u2019s image-input declaration; test the route you will use before relying on screenshots.", "pricingNote": "The Gateway catalog quotes $0.50 input, $0.20 cache reads and $3 output per million, checked October 10. The linked native pricing page returns 404; these are not verified direct-provider list prices.", "accessNote": "The provider identifies a licensed Qwen base without publishing an exact base checkpoint. Weights are not publicly distributed. Hosted and commercial self-hosting access are separate choices; no Qwen benchmark is transferred.", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gpt-6-luna-decisions", "providerId": "gpt-6-luna", "identityLabel": "Decisions API model", "name": "GPT-6 Luna Decisions", "provider": "OpenAI", "released": "2026-10-06", "verification": "Generally available Decisions API", "specialized": "decision", "context": null, "maxOutput": null, "questionLimit": 400, "modalities": [ "text", "image" ], "tools": false, "sourceUrl": "https://developers.openai.com/api/docs/guides/decisions", "officialUrl": "https://developers.openai.com/api/docs/changelog", "documentedRates": { "input": 0.1, "output": 0, "cached": 0.01, "write": 0.125, "write12h": 0.25, "label": "Direct Decisions API \u00b7 USD" }, "note": "Answer yes/no questions, choose named options or score records against your rules. Shared documents can be reused across requests, and each question can include its own images.", "outputType": "Typed predicates, choices and rubric scores with probabilities and confidence fields.", "capabilityNote": "Generally available October 22. Send 1\u2013400 questions with text, images or structured JSON. Public image URLs and base64 are supported; file_id is not. Question images share the request image limit. Context and effort limits are not established by this guide.", "contextNote": "At least 1,024 identical shared-input tokens are needed for caching. Question text and question-specific images are outside the reusable prefix. Cache reads/writes are already included in reported total input tokens; do not add them again.", "pricingNote": "Per million tokens: $0.10 uncached input, $0.01 cache reads, $0.125 standard cache writes or $0.25 for 12-hour writes. Answers are free. Region and long-context premiums still apply.", "billingNote": "This estimate treats input as uncached. Cache reads cost less, cache writes cost more, and regional or long-context premiums can apply.", "accessNote": "A generally available service using GPT-6 Luna, not a new base model. Chat benchmark results and vendor speed claims do not establish decision accuracy or measured latency here.", "discovered": "2026-10-07", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "astabrief-8b", "providerId": "allenai/AstaBrief_8B", "identityLabel": "checkpoint", "name": "AstaBrief 8B", "provider": "Ai2", "released": "2026-10-02", "verification": "Open-weight research model", "specialized": "research", "context": 40960, "maxOutput": null, "modalities": [ "text" ], "tools": null, "sourceUrl": "https://huggingface.co/allenai/AstaBrief_8B", "officialUrl": "https://allenai.org/blog/astabrief", "note": "A model for turning a research question and supplied literature excerpts into a cited report. Its published development studies do not compare today\u2019s frontier models.", "capabilityNote": "Use the recommended report prompt and provide retrieved literature. This checkpoint is a report writer; its upstream retrieval pipeline is separate. The configuration sets max_position_embeddings to 40,960, not a verified hosted service limit.", "contextNote": "40,960 is checkpoint configuration evidence. The example\u2019s 4,096 max_tokens is a request setting, not a published maximum output allowance.", "accessNote": "The checkpoint declares Apache 2.0 and can be run locally. Hosted per-token prices are not verified. Vendor pipeline timings and 2025 development comparisons are not imported into general quality or speed rankings.", "discovered": "2026-10-07", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "kolibri-1", "providerId": "Aleph-Alpha/Kolibri-1", "identityLabel": "checkpoint", "name": "Kolibri 1", "provider": "Aleph Alpha", "released": "2026-10-03", "verification": "Open-weight checkpoint", "context": 1048576, "maxOutput": null, "modalities": [ "text" ], "tools": true, "supportedEfforts": [ "none", "low", "medium", "high" ], "sourceUrl": "https://huggingface.co/Aleph-Alpha/Kolibri-1", "officialUrl": "https://aleph-alpha.com/en/news/kolibri-sovereign-ai-made-in-germany/", "note": "An open-weight German and English reasoning model with tool support. Downloadable weights are available; matched independent general-task scores and hosted token prices are pending.", "contextNote": "The card states 262,144 native context tokens and validates extension to 1,048,576. It recommends at most 262,144 for complex tasks and serving efficiency. These are checkpoint claims, not a hosted API guarantee.", "capabilityNote": "This exact FP8 checkpoint supports text and tool calls through its serving parser. Thinking levels are configured through the chat template; an omitted default is unverified.", "accessNote": "The checkpoint declares Apache 2.0. Running it yourself requires hardware, serving and operational work; it does not establish a zero-cost hosted API.", "discovered": "2026-10-07", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "reka-rho-1", "providerId": null, "name": "Reka Rho-1", "provider": "Reka", "announced": "2026-10-05", "released": null, "verification": "Research preview", "context": null, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": null, "sourceUrl": "https://reka.ai/news/rho-1-collapsing-the-multimodal-stack", "officialUrl": "https://reka.ai/labs", "note": "A research preview that combines text, image and video understanding and generation. General task quality, latency and billing are not independently verified here.", "capabilityNote": "The provider demonstrates one multimodal model and a shared state. The demonstration is not an API tool-calling contract. Public API ID, context and output limits, effort controls and prices remain unverified.", "accessNote": "The provider offers research collaboration. A public announcement does not establish ordinary API-account access.", "discovered": "2026-10-07", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "reflection-beam", "providerId": "Beam-1.0-501B-A23B", "name": "Reflection Beam", "provider": "Reflection", "released": "2026-10-05", "verification": "Weights and hosted access available", "context": 1048576, "maxOutput": 131072, "modalities": [ "text" ], "tools": true, "supportedEfforts": [ "low", "medium", "high", "xhigh", "max" ], "defaultEffort": null, "sourceUrl": "https://developers.reflection.ai/models", "officialUrl": "https://reflection.ai/blog/introducing-beam", "note": "A text model for coding and multi-step work with tools. You can now download it under Apache 2.0 or use a hosted service. Choose hosting for a simpler start; running the full model yourself requires substantial hardware.", "accessNote": "Reflection released weights on October 16 under Apache 2.0. Gateway\u2019s reflection/beam is temporarily free when served by Reflection and automatically becomes paid when that offer ends. reflection/beam-free instead stops serving at the end. Check the selected provider and current terms.", "capabilityNote": "Beam supports tools and structured answers, with low, medium, high, xhigh and max thinking. None and minimal are unsupported. Reflection documents defaults that can differ by organization; check your account\u2019s model response. The documented API name is now Beam-1.0-501B-A23B; this documentation change alone does not establish a new checkpoint or improved quality.", "pricingNote": "Gateway\u2019s Reflection-served offer is temporarily free; neither quote sets your direct Reflection account price. Check the chosen service before budgeting for ongoing use.", "discovered": "2026-10-07", "contextNote": "Reflection documents 1,048,576 combined input/output tokens and up to 131,072 output tokens. Your plan may allow less. Requests that ask for token probabilities (logprobs) allow at most 8,192 output tokens, including thinking. Set an output budget that fits alongside your prompt.", "catalogOffer": { "id": "reflection/beam-1.0-501b-a23b", "input": 0.25, "output": 1.0, "cached": 0.05, "checked": "2026-11-03", "observed": "2026-11-03", "note": "Current OpenRouter rates are shown separately. Use the prices for the service your application calls.", "sourceUrl": "https://openrouter.ai/reflection/beam-1.0-501b-a23b" }, "license": "Apache 2.0", "dateNote": "Announced October 5; weights and broader hosted access became available October 16.", "apiDocumentationUrls": [ "https://reflection.ai/blog/beam-is-available", "https://huggingface.co/reflection/Beam-1.0-501B-A23B", "https://vercel.com/changelog/reflection-ai-beam-now-available-on-ai-gateway-for-free" ], "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "gpt-rosalind-research", "providerId": "gpt-rosalind-research", "name": "GPT-Rosalind Research", "provider": "OpenAI", "discovered": "2026-10-05", "released": null, "verification": "Restricted research access", "specialized": "research", "context": null, "maxOutput": null, "modalities": [], "tools": null, "sourceUrl": "https://developers.openai.com/api/docs/pricing#specialized-models", "officialUrl": "https://developers.openai.com/api/docs/models", "note": "A specialized life-sciences reasoning model for approved research organizations. OpenAI lists October 5 as its billing start; this is a pricing milestone, not a new model launch. Comparable independent benchmarks are pending.", "documentedRates": { "input": 5, "output": 25, "cached": 0.5, "label": "Direct restricted API \u00b7 USD" }, "accessNote": "OpenAI\u2019s September 8 access notice requires approval through its trusted-access program and limits use to approved internal research. A standard API account does not establish access.", "capabilityNote": "OpenAI describes this as a life-sciences reasoning model. The reviewed sources do not establish its context window, output limit, input formats, tools or effort settings.", "pricingNote": "Published standard rates per million tokens: $5 input, $0.50 cached input and $25 output. Billing begins October 5, 2026; an exact cutoff time and timezone are not published. OpenAI says cache-write pricing does not apply. Task costs and measured speed are unavailable.", "dateNote": "Approved research access was announced September 8, 2026. The original model launch date is unverified.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "jev", "providerId": "typesafe-ai/jev", "name": "Jev", "provider": "TypeSafe AI", "verification": "Hosted decision route", "identityLabel": "AI Gateway route", "discovered": "2026-10-04", "released": null, "announced": "2026-09-16", "context": 32000, "maxOutput": null, "modalities": [ "text" ], "tools": false, "specialized": "decision", "questionLimit": null, "outputType": "Typed choices, rubric scores and Boolean probabilities", "sourceUrl": "https://vercel.com/ai-gateway/models/jev", "officialUrl": "https://docs.typesafe.ai/introduction", "capabilitySourceUrl": "https://ai-gateway.vercel.sh/v1/models", "note": "Choose a label, score or yes/no probability for a support ticket, document or other text. Jev is a decision service rather than a chatbot.", "contextNote": "Gateway declares 32,000 tokens for shared state plus the longest question, and 64,000 total request tokens. This is not a 64K single-state context. The maximum question count is unverified.", "capabilityNote": "The exact evaluation contract accepts text state and typed Choice, Score and Boolean questions. It does not stream free-form output; no independent speed advantage or calibrated confidence is inferred.", "pricingNote": "Gateway quotes $0.042 per million billed input tokens and no output-token charge. Its model page rounds the input quote to $0.04; this calculator uses the exact catalog rate.", "documentedRates": { "input": 0.042, "output": 0, "cached": null, "label": "AI Gateway decision offer \u00b7 USD" }, "billingNote": "Gateway explicitly quotes zero output-token price; cache and other fees are not verified.", "accessNote": "Native jev-latest and jev-preview both select 1.14.0 at this check. Use an exact version for repeatable tests. These native aliases do not establish which version the unversioned Gateway offer serves. Gateway uses the fixed-choice evaluation endpoint rather than free-form chat.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "deepseek-v4-1-flash-fast", "providerId": "deepseek/deepseek-v4.1-flash-fast", "name": "DeepSeek V4.1 Flash Fast", "provider": "DeepSeek", "verification": "Hosted serving variant", "identityLabel": "AI Gateway route", "discovered": "2026-10-03", "released": null, "context": 1048576, "maxOutput": 32768, "modalities": [ "text", "image" ], "tools": true, "sourceUrl": "https://vercel.com/ai-gateway/models/deepseek-v4.1-flash-fast", "capabilitySourceUrl": "https://ai-gateway.vercel.sh/v1/models", "gatewayEffortOptions": [ "none", "low", "medium", "high" ], "note": "AI Gateway lists a speed-optimized V4.1 Flash deployment. This is a serving variant; no base-model scores or vendor speed claims become measured results here.", "capabilityNote": "The current Gateway route accepts text and images, supports tools and lists 1,048,576 context with 32,768 maximum output tokens. Its none/low/medium/high options are service controls; no matched effort study is available.", "contextNote": "Limits belong to this hosted route and may differ from another deployment of V4.1 Flash.", "pricingNote": "Gateway\u2019s October13 quote is $0.60 input, $2.40 output and $0.014 cache read per million tokens. These are service-specific rates, separate from DeepSeek\u2019s native price schedule.", "documentedRates": { "input": 0.6, "output": 2.4, "cached": 0.014, "label": "AI Gateway routed offer \u00b7 USD" }, "accessNote": "The exact route has returned to the Gateway catalog after its October9 absence. Its current output allowance and prices differ from the earlier offer; recheck your request limits before switching it back on.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "laya", "providerId": "convaiinnovations/laya", "name": "Laya", "provider": "Convai Innovations", "verification": "Hosted decision route", "identityLabel": "AI Gateway route", "discovered": "2026-10-03", "released": null, "context": 8192, "maxOutput": null, "modalities": [ "text" ], "tools": false, "specialized": "decision", "questionLimit": null, "outputType": "Yes/no probabilities, bounded choices and rubric scores", "sourceUrl": "https://vercel.com/ai-gateway/models/laya", "officialUrl": "https://vercel.com/changelog/laya-decision-model-now-available-on-ai-gateway-free-through-october-31", "note": "Route records, choose labels or score a condition using answers you define. Laya is a paid decision service on Gateway; the free promotion has ended.", "contextNote": "The exact Gateway API lists 8,192 tokens. A maximum question count is unverified; this is not a guarantee that every state plus question fits.", "capabilityNote": "The hosted evaluation API documents text state and typed yes/no, choice and score questions. No hosted image support or independent calibration sweep is established.", "pricingNote": "The current exact Gateway quote is $0.0105 per million billed input tokens, with no output-token charge. The model page rounds this to $0.01. The October 31 free promotion has ended; caching and other fees remain unverified.", "documentedRates": { "input": 0.0105, "output": 0, "cached": null, "label": "Boundless via AI Gateway \u00b7 USD" }, "accessNote": "Use convaiinnovations/laya. The separate convaiinnovations/laya-free endpoint is absent from the current Gateway catalog after the announced promotion ended.", "billingNote": "Use actual billed input. At this rate, 100 million billed tokens cost $1.05; the old zero quote is historical.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "cloudflare-clef-flash", "providerId": "@cf/cloudflare/clef-flash", "name": "Cloudflare Clef Flash", "provider": "Cloudflare", "discovered": "2026-10-02", "released": "2026-10-01", "context": 24576, "contextNote": "Hosted context is 24,576 tokens, including media and questions. Media that exceeds the window makes a request fail; otherwise text state is truncated to fit the remaining space. Inspect retained context before acting on a decision.", "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": false, "specialized": "decision", "questionLimit": 64, "outputType": "Yes/no probabilities, bounded choices and rubric scores", "sourceUrl": "https://developers.cloudflare.com/workers-ai/models/clef-flash/", "officialUrl": "https://blog.cloudflare.com/clef-decision-models/", "note": "Provider-announced multimodal decision model with Apache 2.0 weights. Workers AI exposes bounded typed questions, not a matched chat-model result. Vendor quality/latency comparisons have not been independently reproduced here.", "capabilityNote": "The model card declares text, JSON, images and video input. The hosted schema explicitly documents 1\u201364 named questions and up to four embedded images; hosted video request support is not established by the model-card claim. No adjustable effort sweep is verified.", "pricingNote": "The exact Workers AI page lists $0.038 per million billed input tokens and explicitly no output-token charge. Images add input tokens. A cache-read price and any other charges remain unverified; use actual billed usage.", "documentedRates": { "input": 0.038, "output": 0, "cached": null, "label": "Workers AI documented rates \u00b7 USD" }, "accessNote": "Hosted through Workers AI using a Cloudflare account and API authorization; Apache 2.0 weights can be run separately. Hosting costs and local inference costs differ.", "billingNote": "Workers AI explicitly documents free output tokens. Media tokens are billed at the input rate; other fees and cache rates remain unverified.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "cloudflare-clef", "providerId": "@cf/cloudflare/clef", "name": "Cloudflare Clef", "provider": "Cloudflare", "discovered": "2026-10-02", "released": "2026-10-01", "context": 65536, "contextNote": "Hosted context is 65,536 tokens, including media and questions. Media that exceeds the window makes a request fail; otherwise text state is truncated to fit the remaining space. Inspect retained context before acting on a decision.", "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": false, "specialized": "decision", "questionLimit": 64, "outputType": "Yes/no probabilities, bounded choices and rubric scores", "sourceUrl": "https://developers.cloudflare.com/workers-ai/models/clef/", "officialUrl": "https://blog.cloudflare.com/clef-decision-models/", "note": "Provider-announced multimodal decision model with Apache 2.0 weights. Workers AI exposes bounded typed questions, not a matched chat-model result. Vendor quality/latency comparisons have not been independently reproduced here.", "capabilityNote": "The model card declares text, JSON, images and video input. The hosted schema explicitly documents 1\u201364 named questions and up to four embedded images; hosted video request support is not established by the model-card claim. No adjustable effort sweep is verified.", "pricingNote": "The exact Workers AI page lists $0.24 per million billed input tokens and explicitly no output-token charge. Images add input tokens. A cache-read price and any other charges remain unverified; use actual billed usage.", "documentedRates": { "input": 0.24, "output": 0, "cached": null, "label": "Workers AI documented rates \u00b7 USD" }, "accessNote": "Hosted through Workers AI using a Cloudflare account and API authorization; Apache 2.0 weights can be run separately. Hosting costs and local inference costs differ.", "billingNote": "Workers AI explicitly documents free output tokens. Media tokens are billed at the input rate; other fees and cache rates remain unverified.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "perplexity-decider-v1-27b", "providerId": "pplx-decider-v1-27b", "name": "Perplexity Decider v1 27B", "provider": "Perplexity", "discovered": "2026-10-02", "released": null, "context": 262144, "contextNote": "The native limit is fewer than 256K tokens for the shared information, images and each question. Large multi-question requests can be split. Up to 128 questions, 255 choice options, 10 score levels and a 32 MiB request body. Gateway\u2019s 262,144 declaration is separate.", "maxOutput": null, "modalities": [ "text", "image" ], "tools": false, "specialized": "decision", "questionLimit": 128, "outputType": "Yes/no probabilities, bounded choices and rubric scores", "sourceUrl": "https://docs.perplexity.ai/docs/decisions/quickstart", "officialUrl": "https://huggingface.co/perplexity-ai/pplx-decider-v1-27b", "note": "This older API name now runs Decider v1.2. Use it to classify records or choose among answers you supply, but recheck any rules that rely on the model\u2019s probabilities after the version change.", "capabilityNote": "The current API accepts text and images and up to 128 named questions. Answers are probabilities, choices or ordered scores. Grouping questions can change the answers; keep the grouping consistent when testing.", "pricingNote": "Requests using this older name now cost $0.02 per million input tokens, with free answers and no per-request fee. Related questions can share the input charge; splitting a large request counts the shared information again.", "documentedRates": { "input": 0.02, "output": 0, "cached": null, "label": "Direct Decisions API \u00b7 USD" }, "accessNote": "Perplexity redirects pplx-decider-v1-27b and pplx-decider-v1.1-27b to pplx-decider-v1.2. The response repeats the name you sent, so it does not prove that the old model ran. Downloaded v1 weights remain a separate model.", "verification": "Older API name now selects v1.2", "billingNote": "Current native requests run v1.2. Count the billed input reported by the service; grouped questions share state, while split groups count it again.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "ling-3-1-flash", "providerId": "inclusionai/ling-3.1-flash", "name": "Ling 3.1 Flash", "provider": "InclusionAI", "verification": "Hosted route documented", "identityLabel": "AI Gateway route", "discovered": "2026-10-02", "released": null, "context": 262144, "contextNote": "AI Gateway\u2019s exact models API declares a 262,144-token hosted context and 32,768-token output allowance; the model webpage rounds context to 262K. Gateway availability was announced September 30, separately from the unverified base-model launch date.", "maxOutput": 32768, "modalities": [ "text" ], "tools": true, "sourceUrl": "https://vercel.com/ai-gateway/models/ling-3.1-flash", "officialUrl": "https://vercel.com/changelog/ling-3-1-flash-is-now-available-on-ai-gateway", "note": "A text model for coding, analysis and tools. The Gateway free promotion has ended; its standard offer now has published token prices.", "accessNote": "The standard Gateway route is now paid. The separate -free route is absent from the October 14 catalog. This does not establish a native-provider shutdown.", "pricingNote": "Gateway now quotes $0.30 input, $0.06 cached input and $0.90 output per million tokens. Earlier free quotes apply to the ended promotion, not this current offer.", "documentedRates": { "input": 0.3, "output": 0.9, "cached": 0.06, "label": "AI Gateway paid offer \u00b7 USD" }, "capabilityNote": "Gateway supports text, tools and a 32,768-token answer budget that includes thinking. Its thinking controls are listed below; comparable independent results and the default setting are unverified.", "gatewayEffortOptions": [ "none", "low", "medium", "high" ], "capabilitySourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogOffer": { "id": "inclusionai/ling-3.1-flash", "sourceUrl": "https://openrouter.ai/inclusionai/ling-3.1-flash", "observed": "2026-11-03", "input": 0.075, "output": 0.225, "cached": 0.015, "note": "Current OpenRouter rates are shown separately. Use the prices for the service your application calls.", "checked": "2026-11-03" }, "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "minimax-m3-1-flash-preview", "providerId": "MiniMax-M3.1-Flash-Preview", "name": "MiniMax M3.1 Flash Preview", "provider": "MiniMax", "discovered": "2026-10-01", "context": 1000000, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": true, "supportedEfforts": [ "low", "medium", "high", "xhigh", "max" ], "sourceUrl": "https://www.minimax.io/m-plan", "officialUrl": "https://agent.minimax.io/tools/m3-1-flash-preview", "note": "The earlier M Plan and MiniMax Code preview. The stable M3.1 Flash release has its own page; do not use this preview\u2019s identity for new tests.", "defaultEffort": "max", "apiDocumentationUrls": [ "https://platform.minimax.io/docs/api-reference/text-openai-api", "https://platform.minimax.io/docs/api-reference/text-anthropic-api", "https://www.minimax.io/blog/minimax-m31f" ], "accessNote": "MiniMax\u2019s current API documentation now names MiniMax-M3.1-Flash. This historical preview page retains the previously documented settings; the preview ID\u2019s present behavior is not verified.", "maxOutputNote": "A provider maximum output allowance is not verified. The 4,096-token request in the Anthropic documentation is an example, not a published model limit; thinking and answer tokens share the request budget.", "checked": "2026-11-03", "catalogId": null, "catalogAdded": null, "released": null, "verification": "Provider documented", "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "slug": "particle-1", "catalogId": "particle-ai/particle-1", "name": "Particle-1", "provider": "Particle AI", "released": "2026-10-28", "catalogAdded": "2026-10-28", "checked": "2026-11-03", "verification": "Provider and hosting offer verified", "note": "Try a coding model trained to finish software tasks with fewer generated tokens. Compare finished work, retries and the whole bill on your own repositories.", "context": 1048576, "maxOutput": 131072, "modalities": [ "text", "image" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/particle-ai/particle-1", "officialUrl": "https://particle.ai/blog/introducing-particle-1", "rates": { "input": 0.05, "output": 0.5, "cached": 0.003, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilityNote": "Adapted from DeepSeek V4.1 Flash for coding and data-engineering tasks. It accepts text and images and can call tools. Publisher efficiency tests do not establish an independent quality score or guaranteed savings.", "contextNote": "Native and routed documentation list 1,048,576 tokens shared by input and output, with a 131,072 output ceiling. Native requests default to 32,000 output tokens when no limit is supplied. Thinking uses the same output budget.", "accessNote": "The native request alias is particle-1; responses identify particle-ai/particle-1. Native thinking defaults to high: none disables it; minimal/low map to low, medium/high/xhigh to high, and max/ultra to max. Gateway lists none/low/high/max separately. A separate US hostname restricts inference GPUs to the United States. It does not restrict file storage, billing records or other services; retries must stay on the US hostname.", "pricingNote": "Native standard rates are $0.25 input, $0.03 cached input and $1 output per million tokens. Off-peak rates are half, from 20:30 to 04:00 America/New_York. Effective Flex halves token prices again; requesting Flex alone does not guarantee entitlement. Web searches cost extra. Router and Gateway quotes apply to their own offers.", "apiDocumentationUrls": [ "https://docs.particle.ai/models.md", "https://docs.particle.ai/reasoning.md", "https://docs.particle.ai/usage-billing.md", "https://docs.particle.ai/service-tiers.md", "https://docs.particle.ai/regional-endpoints.md" ], "independentEvidence": [] }, { "slug": "ternary-bonsai-2-9b", "catalogId": "prism-ml/ternary-bonsai-2-9b", "name": "Ternary Bonsai 2 9B", "provider": "PrismML", "released": null, "catalogAdded": "2026-10-16", "checked": "2026-11-03", "verification": "Provider and hosting offer verified", "note": "Use a compact text-and-image model through OpenRouter, or download it for a compatible local setup. Its compressed files reduce storage needs; test quality and total memory on your own task.", "context": 262144, "maxOutput": 65536, "modalities": [ "text", "image" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/prism-ml/ternary-bonsai-2-9b", "officialUrl": "https://huggingface.co/prism-ml/Ternary-Bonsai-2-9B-gguf", "rates": { "input": 0.08, "output": 0.13, "cached": 0.04, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilityNote": "Derived from Qwen3.5-9B, with ternary weights. This model has no independent result here and does not inherit Qwen\u2019s scores.", "contextNote": "OpenRouter lists 262,144 context tokens and up to 65,536 output tokens. The download also documents 262,144 context; its separate output maximum is not established.", "accessNote": "Apache 2.0 weights are downloadable. Use PrismML\u2019s llama.cpp fork for its supported compressed formats; images require the separate vision file. File size is not total runtime memory.", "apiDocumentationUrls": [ "https://huggingface.co/prism-ml/Ternary-Bonsai-2-9B-gguf/blob/main/LICENSE" ], "independentEvidence": [] }, { "slug": "gpt-6-1-terra-pro", "catalogId": "openai/gpt-6.1-terra-pro", "name": "GPT-6.1 Terra Pro \u00b7 OpenRouter", "provider": "OpenAI", "released": null, "catalogAdded": "2026-10-26", "checked": "2026-11-03", "verification": "Catalog variant", "note": "OpenRouter serves the Terra model with extra reasoning in pro mode. It can use substantially more tokens; this is a hosting option, not a separately verified base model.", "context": 1050000, "maxOutput": 128000, "modalities": [ "file", "image", "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/openai/gpt-6.1-terra-pro", "officialUrl": null, "rates": { "input": 1.0, "output": 5.0, "cached": 0.05, "write": 1.25, "overrides": [ { "min_prompt_tokens": 272000, "prompt": 2.0, "completion": 7.5, "input_cache_read": 0.1, "input_cache_write": 2.5 } ] }, "minimumCacheTokens": null, "capabilityNote": "The catalog describes pro reasoning mode on the same Terra model. Do not treat its higher token use as a verified quality improvement.", "accessNote": "No separate native OpenAI Terra Pro model page was found. This page describes the exact OpenRouter offer; scores from Terra or other GPT models are not assigned to it.", "independentEvidence": [] }, { "slug": "gpt-6-1-terra", "catalogId": "openai/gpt-6.1-terra", "name": "GPT-6.1 Terra", "provider": "OpenAI", "released": "2026-10-26", "catalogAdded": "2026-10-26", "checked": "2026-11-03", "verification": "Provider verified", "note": "Try Terra for reading screenshots, comparing product images, working in apps or revising text. Its standard token prices are half Sol\u2019s; compare the results on your own tasks before switching.", "context": 1050000, "maxOutput": 128000, "modalities": [ "file", "image", "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/openai/gpt-6.1-terra", "officialUrl": "https://developers.openai.com/api/docs/models/gpt-6.1-terra", "rates": { "input": 1.0, "output": 5.0, "cached": 0.05, "write": 1.25, "overrides": [ { "min_prompt_tokens": 272000, "prompt": 2.0, "completion": 7.5, "input_cache_read": 0.1, "input_cache_write": 2.5 } ] }, "minimumCacheTokens": null, "supportedEfforts": [ "none", "low", "medium", "high", "xhigh", "max" ], "defaultEffort": "medium", "effortScope": "OpenAI native API", "contextNote": "The 1,050,000-token combined budget allows at most 922,000 input and 128,000 output tokens. Thinking and the answer share the output allowance.", "capabilityNote": "Use gpt-6.1-terra for OpenAI\u2019s native API. Accepts text and images and returns text. Tool calling requires the Responses API; Chat Completions supports requests without tools. Audio and video input are not supported.", "pricingNote": "Native Standard costs $1 input, $0.05 cached input and $5 output per million tokens. Above 272,000 input tokens, the whole request costs twice the input/cache rate and 1.5 times the output rate. Cache writes cost $1.25 for 30-minute retention or $2.50 for 12-hour retention, before the long-input adjustment.", "accessNote": "Fast costs twice Standard; Batch and Flex cost half Standard. These are service options for the same model. No matched independent result or measured response time is available here.", "apiDocumentationUrls": [ "https://developers.openai.com/api/docs/guides/latest-model", "https://developers.openai.com/api/docs/pricing" ], "independentEvidence": [] }, { "slug": "lfm-2-5-2-6b", "catalogId": "liquid/lfm-2.5-2.6b", "name": "LFM2.5 2.6B \u00b7 OpenRouter", "provider": "Liquid AI", "released": null, "catalogAdded": "2026-08-11", "checked": "2026-11-03", "verification": "Hosted offer verified", "note": "A small reasoning model for retrieval, extraction and tools. The publisher does not recommend it for agentic coding or knowledge-heavy tasks. This is newly discovered hosting, not a newly dated model release.", "context": 128000, "maxOutput": 8192, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/liquid/lfm-2.5-2.6b", "officialUrl": "https://huggingface.co/LiquidAI/LFM2.5-2.6B", "rates": { "input": 0.35, "output": 0.5, "cached": 0.04, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilityNote": "Reasoning is always enabled. Named effort levels and a default effort are not established.", "contextNote": "OpenRouter lists 128,000 context and 8,192 maximum output tokens. The separate downloadable checkpoint declares 131,072 context.", "accessNote": "Downloadable weights use the LFM Open License; review its commercial-use conditions. Hosting access and data terms are separate.", "pricingNote": "This offer lists $0.35 input, $0.04 cached input and $0.50 output per million tokens.", "independentEvidence": [] }, { "slug": "kimi-k3-1-pro", "catalogId": "moonshotai/kimi-k3.1-pro", "name": "Kimi K3.1 Pro", "provider": "Moonshot AI", "released": "2026-10-23", "catalogAdded": "2026-10-23", "checked": "2026-11-03", "verification": "Hosted offer verified", "note": "Try the K3 update for coding and long-document work. Compare it on your own tasks before replacing K3; independent results are still pending.", "context": 1048576, "maxOutput": 943718, "modalities": [ "text", "image", "video" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/moonshotai/kimi-k3.1-pro", "officialUrl": "https://platform.kimi.ai/docs/guide/kimi-k3-1-pro-quickstart", "rates": { "input": 2.0, "output": 10.0, "cached": 0.2, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilityNote": "Native Kimi always thinks, with low/high/max effort and max by default. OpenRouter separately lists a none option. Check the service you call instead of assuming the settings are interchangeable.", "accessNote": "Native access requires a successful top-up of at least $1. The publisher promises weights by November 4; a download is not yet verified.", "contextNote": "This OpenRouter offer declares 1,048,576 context and 943,718 maximum output tokens. Thinking and the answer share the output budget.", "pricingNote": "OpenRouter lists $2 input, $0.20 cached input and $10 output per million tokens. These are hosted prices; a readable native price table was unavailable. Native batch processing is available at 60% of standard price for work that does not need an immediate response. This does not change the routed rate used in the calculator.", "defaultEffort": "max", "supportedEfforts": [ "none", "low", "high", "max" ], "effortScope": "OpenRouter", "independentEvidence": [] }, { "slug": "flint-v1-0-free", "catalogId": "recursive/flint-v1.0:free", "name": "Recursive Flint \u00b7 OpenRouter free", "provider": "Recursive", "released": null, "catalogAdded": "2026-10-21", "checked": "2026-11-03", "verification": "New hosted model verified", "note": "Try Flint for coding and tool-assisted text tasks through a currently free OpenRouter offer. Check availability before building around a free service.", "context": 262144, "maxOutput": 235929, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/recursive/flint-v1.0:free", "officialUrl": "https://huggingface.co/RecursiveSI/flint", "rates": { "input": 0.0, "output": 0.0, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilityNote": "Reasoning is required; the catalog does not establish named effort levels or a default effort. Your application runs the tools.", "accessNote": "The free hosting offer has its own limits. Apache 2.0 weights are separately downloadable; hosting terms and data handling are separate.", "contextNote": "OpenRouter lists 262,144 context and 235,929 maximum output tokens. Thinking and the answer share the output budget.", "pricingNote": "Input and output are currently quoted at zero. No expiry date or permanent free-access promise is verified.", "independentEvidence": [] }, { "slug": "mercury-3", "catalogId": "inception/mercury-3", "name": "Mercury 3", "provider": "Inception", "released": "2026-10-19", "catalogAdded": "2026-10-16", "checked": "2026-11-03", "verification": "Catalog verified", "note": "A text model designed for responsive chat and tool-driven tasks. Compare it on your own examples before choosing it for a speed-sensitive workflow.", "context": 1000000, "maxOutput": 131072, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/inception/mercury-3", "officialUrl": "https://www.inceptionlabs.ai/blog/introducing-mercury-3", "rates": { "input": 0.1, "output": 0.35, "cached": 0.01, "write": null, "overrides": [] }, "minimumCacheTokens": null, "providerId": "mercury-3", "supportedEfforts": [ "none", "low", "medium", "high" ], "defaultEffort": "medium", "capabilityNote": "Text input and output, tool calling and structured answers. OpenRouter lists none/low/medium/high and defaults to medium. The native API instead names its fastest setting instant and defaults to low; choose the setting explicitly when comparing services.", "contextNote": "Native, Gateway and OpenRouter list a 1,000,000-token context and up to 131,072 output tokens.", "pricingNote": "The current hosted quote matches the native launch discount: $0.10 input, $0.01 cached input and $0.35 output per million tokens. Native regular rates are $0.20/$0.02/$0.70; no end date for the discount is verified.", "accessNote": "Available through Inception and hosted providers. Native instant/low/medium/high options do not establish equivalent behavior to OpenRouter\u2019s none setting. Vendor speed claims and partner testimonials are not matched independent measurements.", "apiDocumentationUrls": [ "https://docs.inceptionlabs.ai/get-started/models", "https://docs.inceptionlabs.ai/capabilities/reasoning-efforts" ], "effortScope": "OpenRouter", "independentEvidence": [] }, { "slug": "minimax-m3-1-flash-free", "catalogId": "minimax/minimax-m3.1-flash:free", "name": "MiniMax M3.1 Flash \u00b7 OpenRouter free", "provider": "MiniMax", "released": "2026-10-19", "catalogAdded": "2026-10-19", "checked": "2026-11-03", "verification": "Free offer no longer listed", "note": "This free OpenRouter offer is no longer listed. The stable model is available through a separate paid offer.", "context": null, "maxOutput": null, "modalities": [ "text", "image", "video" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/minimax/minimax-m3.1-flash:free", "officialUrl": "https://www.minimax.io/blog/minimax-m31f", "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "supportedEfforts": [ "low", "medium", "high", "xhigh", "max" ], "defaultEffort": "max", "capabilityNote": "Before removal, OpenRouter listed text, image and video input, tools, and always-on thinking. These are historical features of the removed offer.", "contextNote": "The former free offer listed 1,000,000 context tokens and 524,288 maximum output tokens. No current capacity is verified for this removed ID.", "pricingNote": "The former zero-price quote is historical. Current access and prices for this exact free ID are not verified.", "accessNote": "The free OpenRouter ID is absent on October 23. Gateway\u2019s MiniMax-served promotion separately ended October 22; do not assume either service is still free.", "apiDocumentationUrls": [ "https://vercel.com/changelog/minimax-m3-1-flash-now-available-on-ai-gateway-for-free" ], "effortScope": "OpenRouter", "independentEvidence": [], "availability": "unlisted" }, { "slug": "pareto-26-10-openrouter", "catalogId": "unbiased/pareto-26.10", "name": "Pareto 26.10 \u00b7 OpenRouter", "provider": "Unbiased", "released": null, "catalogAdded": "2026-10-15", "checked": "2026-11-03", "verification": "Catalog verified", "note": "A hosted option for the stable October service, which combines several models into one answer. Its token prices match the published native quote; access and data handling still depend on the service you choose.", "context": 1048576, "maxOutput": 131072, "modalities": [ "text", "image" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/unbiased/pareto-26.10", "officialUrl": "https://unbiased.ai/blog/pareto-26-10/", "rates": { "input": 0.8, "output": 3.2, "cached": 0.03, "write": null, "overrides": [] }, "minimumCacheTokens": null, "contextNote": "OpenRouter lists 1,048,576 context tokens and 131,072 maximum output tokens. The native provider separately advertises one-million-token context without an exact output limit here.", "dateNote": "The native stable release was announced October 15. This newly observed OpenRouter offer is separate from the retired preview.", "capabilityNote": "This is a service that combines models, not a single downloadable model. Earlier preview evaluations are not new measurements of the stable version.", "accessNote": "Unbiased now documents this exact OpenRouter ID. OpenRouter handles the bill and its own data rules also apply. The zero-retention promise covers designated API traffic, not Unbiased Chat or Personal subscriptions. This model is separate from OpenRouter\u2019s own Pareto Code Router.", "apiDocumentationUrls": [ "https://unbiased.ai/developers/openrouter", "https://unbiased.ai/terms" ], "independentEvidence": [] }, { "slug": "glimpse", "catalogId": "elorian/glimpse", "name": "Elorian Glimpse \u00b7 OpenRouter", "provider": "Elorian", "released": null, "catalogAdded": "2026-10-17", "checked": "2026-11-03", "verification": "Catalog verified", "note": "Ask questions about charts, images and video. This hosting option has different context and output limits from Gateway; choose the service that fits your workload.", "context": 262144, "maxOutput": 256000, "modalities": [ "text", "image", "video" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/elorian/glimpse", "officialUrl": "https://elorian.ai/blog/introducing-glimpse", "rates": { "input": 0.6, "output": 3.0, "cached": 0.06, "write": null, "overrides": [] }, "minimumCacheTokens": null, "contextNote": "This OpenRouter offer lists 262,144 context tokens and 256,000 maximum output tokens. Gateway separately lists 220,000 and 4,096; these limits are not interchangeable.", "capabilityNote": "OpenRouter lists text, image and video input plus tool calling. It does not enumerate named thinking levels in today\u2019s catalog. Do not assume Gateway\u2019s settings or the vendor\u2019s built-in visual tools are available here.", "dateNote": "Glimpse launched October 16. This is an additional hosting offer, not a new version.", "independentEvidence": [] }, { "slug": "deepseek-v4-1-flash-1017", "catalogId": "deepseek/deepseek-v4.1-flash-1017", "name": "DeepSeek V4.1 Flash 1017", "provider": "DeepSeek", "released": "2026-10-17", "catalogAdded": "2026-10-17", "checked": "2026-11-03", "verification": "Catalog verified", "note": "The October 17 update is available for coding, visual questions and work with tools. DeepSeek reports faster, shorter answers, but independent results for this exact revision are still pending.", "context": 1048576, "maxOutput": 393216, "modalities": [ "text", "image" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/deepseek/deepseek-v4.1-flash-1017", "officialUrl": "https://api-docs.deepseek.com/news/news261017/", "rates": { "input": 0.3, "output": 1.2, "cached": 0.006, "write": null, "overrides": [ { "utc_days": [ "saturday", "sunday" ], "prompt": 0.15, "completion": 0.6, "input_cache_read": 0.003 }, { "utc_days": [ "monday", "tuesday", "wednesday", "thursday", "friday" ], "utc_start": 0, "utc_end": 100, "prompt": 0.15, "completion": 0.6, "input_cache_read": 0.003 }, { "utc_days": [ "monday", "tuesday", "wednesday", "thursday", "friday" ], "utc_start": 100, "utc_end": 400, "prompt": 0.3, "completion": 1.2, "input_cache_read": 0.006 }, { "utc_days": [ "monday", "tuesday", "wednesday", "thursday", "friday" ], "utc_start": 400, "utc_end": 600, "prompt": 0.15, "completion": 0.6, "input_cache_read": 0.003 }, { "utc_days": [ "monday", "tuesday", "wednesday", "thursday", "friday" ], "utc_start": 600, "utc_end": 1000, "prompt": 0.3, "completion": 1.2, "input_cache_read": 0.006 }, { "utc_days": [ "monday", "tuesday", "wednesday", "thursday", "friday" ], "utc_start": 1000, "utc_end": 0, "prompt": 0.15, "completion": 0.6, "input_cache_read": 0.003 } ] }, "minimumCacheTokens": null, "supportedEfforts": [ "none", "low", "high", "max" ], "defaultEffort": "high", "effortScope": "OpenRouter offer", "license": "MIT", "contextNote": "OpenRouter lists 1,048,576 context tokens and 393,216 maximum output tokens. DeepSeek\u2019s native API and Gateway instead document 1M context and 384K output; use the limits of the service you call.", "accessNote": "DeepSeek\u2019s deepseek-flash name now selects this October 17 revision. The retired deepseek-v4-flash and deepseek-v4-flash-vision-exp names also run it. To compare versions, use a pinned hosting offer or downloaded checkpoint rather than those moving names.", "capabilityNote": "The model accepts text and images, supports tools and can turn thinking on or off. DeepSeek\u2019s own tests use specific tools and settings, so they do not establish the results you will get in your application.", "pricingNote": "The displayed OpenRouter quote is off-peak; its catalog also lists peak rates. Buying directly from DeepSeek costs $0.15 input, $0.003 cached input and $0.60 output per million tokens off-peak; peak prices are double. Peak hours are weekdays 01:00\u201304:00 and 06:00\u201310:00 UTC, excluding Chinese public holidays. DeepSeek\u2019s Search and Fetch each cost $0.75 per 1,000 successful requests; the free period for built-in calls through DeepSeek Harness, Claude Code and Codex ends October 31 at 16:00 UTC.", "additionalSources": [ "https://api-docs.deepseek.com/quick_start/pricing/", "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017", "https://vercel.com/changelog/deepseek-v4-1-flash-1017-now-available-on-ai-gateway" ], "independentEvidence": [] }, { "slug": "xing4-0-29b-a4b-free", "catalogId": "xingchen-agi/xing4.0-29b-a4b:free", "name": "Xing4.0 29B A4B \u00b7 Free offer", "provider": "China Telecom / TeleAI", "released": null, "catalogAdded": "2026-10-16", "checked": "2026-11-03", "verification": "Catalog verified", "note": "An open-weight text model for coding, document questions and tool-based workflows. This newly listed free service is one way to try it; running your own copy still needs hardware. Independent comparison results are pending.", "context": 262144, "maxOutput": 235929, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/xingchen-agi/xing4.0-29b-a4b:free", "officialUrl": "https://huggingface.co/XingChen-AGI/Xing4.0-29B-A4B", "rates": { "input": 0.0, "output": 0.0, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilityNote": "The model supports tool use and thinking that can be enabled or disabled. It has 29 billion total parameters and activates 4 billion at a time; that does not mean the full weights fit in 4 billion parameters\u2019 worth of memory.", "contextNote": "The hosted offer lists 262,144 context tokens and 235,929 maximum output tokens. The downloadable model card describes 256K context, extendable to 512K with a different setup; the extension is not assumed for this hosted offer.", "accessNote": "The publisher provides Apache 2.0 weights and code. The OpenRouter free offer has no stated expiry in today\u2019s catalog; availability can change.", "dateNote": "This is a newly discovered hosted offer for an already published checkpoint. The exact original release day is not established by the accessible model card.", "license": "Apache 2.0", "supportedEfforts": [ "off", "on" ], "effortScope": "self-hosted model", "independentEvidence": [] }, { "slug": "beam-1-0-501b-a23b", "catalogId": "reflection/beam-1.0-501b-a23b", "name": "Reflection Beam 1.0 \u00b7 OpenRouter", "provider": "Reflection", "released": "2026-10-16", "catalogAdded": "2026-10-16", "checked": "2026-11-03", "verification": "Catalog verified", "note": "A paid hosted offer for Beam\u2019s coding and tool-based workflows. Compare its ongoing cost with Gateway\u2019s temporary free access; neither option has a matched independent quality result here.", "context": 1048576, "maxOutput": 131072, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/reflection/beam-1.0-501b-a23b", "officialUrl": "https://reflection.ai/blog/beam-is-available", "rates": { "input": 0.25, "output": 1.0, "cached": 0.05, "write": null, "overrides": [] }, "minimumCacheTokens": null, "contextNote": "OpenRouter now lists 1,048,576 context and 131,072 maximum output tokens, matching the native standard output allowance. Earlier larger hosted declarations remain historical.", "supportedEfforts": [ "low", "medium", "high", "xhigh", "max" ], "defaultEffort": "max", "effortScope": "OpenRouter offer", "capabilityNote": "The OpenRouter offer defaults to max thinking and cannot turn thinking off. Native account defaults can differ. Vendor benchmarks are not a matched effort study.", "accessNote": "Weights became available October 16. The October 5 announcement and October 16 hosted offers are different milestones, not proof of a new checkpoint.", "independentEvidence": [] }, { "slug": "moondream3-1-9b-a2b", "catalogId": "moondream/moondream3.1-9b-a2b", "name": "Moondream 3.1", "provider": "Moondream", "released": "2026-07-07", "catalogAdded": "2026-10-15", "checked": "2026-11-03", "verification": "Catalog verified", "note": "Describe images, locate objects or answer questions about a picture. This is newly covered on our site; the model itself was released earlier.", "context": 32000, "maxOutput": 4096, "modalities": [ "text", "image" ], "tools": false, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/moondream/moondream3.1-9b-a2b", "officialUrl": "https://moondream.ai/blog/moondream-3-1-beyond-benchmarks", "rates": { "input": 0.3, "output": 2.5, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "discovered": "2026-10-16", "capabilityNote": "The native model has separate skills for captions, questions, points and object boxes. Confirm that your chosen host exposes the skill you need; a chat endpoint may not expose every native feature.", "accessNote": "Moondream 3.1 can run locally with Photon on supported NVIDIA GPUs or Apple Silicon. Its custom model license allows building applications but restricts reselling general-purpose model hosting or fine-tuning.", "pricingNote": "This OpenRouter offer lists $0.30 input and $2.50 output per million tokens. The provider\u2019s July 7 launch lists its own Cloud service at $0.30/$1.00; choose the price for the service you actually use.", "independentEvidence": [] }, { "slug": "moondream3-preview", "catalogId": "moondream/moondream3-preview", "name": "Moondream 3 Preview", "provider": "Moondream", "released": "2025-09-18", "catalogAdded": "2026-10-15", "checked": "2026-11-03", "verification": "Catalog verified", "note": "Describe images, locate objects or answer questions about a picture. This is newly covered on our site; the model itself was released earlier.", "context": 32000, "maxOutput": 4096, "modalities": [ "text", "image" ], "tools": false, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/moondream/moondream3-preview", "officialUrl": "https://moondream.ai/blog/moondream-3-preview", "rates": { "input": 0.3, "output": 2.5, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "discovered": "2026-10-16", "capabilityNote": "The native model has separate skills for captions, questions, points and object boxes. Confirm that your chosen host exposes the skill you need; a chat endpoint may not expose every native feature.", "accessNote": "The preview supports thinking on or off for image questions. Its provider benchmarks are separate from our independent matched results.", "pricingNote": "These prices describe the selected OpenRouter offer, not the cost of running downloaded weights.", "independentEvidence": [] }, { "slug": "nemotron-3-5-super-vl", "catalogId": "nvidia/nemotron-3.5-super-vl", "name": "Nemotron 3.5 Super VL", "provider": "NVIDIA", "released": "2026-10-15", "catalogAdded": "2026-10-15", "checked": "2026-11-03", "verification": "Catalog verified", "note": "Read documents, screenshots and video, then produce text or call tools. This new NVIDIA model has no matched independent score in this comparison yet.", "context": 262144, "maxOutput": 235929, "modalities": [ "text", "image", "video" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/nvidia/nemotron-3.5-super-vl", "officialUrl": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4", "rates": { "input": 0.15, "output": 0.5, "cached": 0.02, "write": null, "overrides": [] }, "minimumCacheTokens": null, "discovered": "2026-10-16", "contextNote": "The downloadable model supports up to one million tokens, but this OpenRouter offer currently lists 262,144 for its serving provider. The calculator uses that smaller limit. Gateway separately lists 1,048,576; confirm your chosen host before sending a large document.", "capabilityNote": "The native model supports thinking off, on with a medium budget, or on. Those controls have not been tested here on the same questions. Do not reuse Lightning or older Nemotron scores for this release.", "accessNote": "Weights are available under OpenMDW 1.1. The separate free OpenRouter route and paid Gateway route have their own limits and availability; a free offer is not a permanent promise for every host.", "independentEvidence": [] }, { "slug": "muse-spark-open", "catalogId": "meta/muse-spark-open", "name": "Muse Spark Open", "provider": "Meta", "released": "2026-10-15", "catalogAdded": "2026-10-15", "checked": "2026-11-03", "verification": "Catalog verified", "note": "Use Meta\u2019s downloadable model for coding and workflows that read text, images, audio or video. A hosted API avoids the large hardware investment needed to run it yourself.", "context": 1048576, "maxOutput": 943718, "modalities": [ "text", "image", "video", "file", "audio" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/meta/muse-spark-open", "officialUrl": "https://dev.meta.ai/docs/muse-spark-open", "rates": { "input": 1.25, "output": 4.25, "cached": 0.15, "write": null, "overrides": [] }, "minimumCacheTokens": null, "discovered": "2026-10-16", "supportedEfforts": [ "low", "medium", "high", "xhigh" ], "defaultEffort": "xhigh", "capabilityNote": "Reads text, images, audio and video and produces text. Tool support lets an application connect it to actions such as searching files or running code; your application supplies those tools.", "accessNote": "This model is not served by Meta Model API. Choose a hosting partner or download it under the Muse Model License. The compressed weights alone occupy about 537 GB; full-precision serving needs roughly 2 TB of GPU memory. The 42B active parameter count does not mean the whole model fits on a small GPU.", "contextNote": "The model supports 1,048,576 combined context tokens. The OpenRouter offer shown here limits generated output to 943,718 tokens; other hosts can set different limits.", "effortScope": "self-hosted model", "independentEvidence": [] }, { "slug": "nemotron-3-5-lightning", "catalogId": "nvidia/nemotron-3.5-lightning", "name": "Nemotron 3.5 Lightning", "provider": "NVIDIA", "released": null, "catalogAdded": "2026-08-11", "checked": "2026-11-03", "verification": "Catalog verified", "note": "A smaller open model for text tasks and tools. LiveBench now includes its own measured result; its exact thinking budget is not specified in the result ID.", "context": 262144, "maxOutput": 235929, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "nemotron-3-5-lightning", "sourceUrl": "https://openrouter.ai/nvidia/nemotron-3.5-lightning", "officialUrl": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16", "rates": { "input": 0.07, "output": 0.2, "cached": 0.04, "write": null, "overrides": [] }, "minimumCacheTokens": null, "discovered": "2026-10-16", "capabilityNote": "The new score belongs to the exact 30B A3B Lightning model in LiveBench. It does not establish the performance of Super VL, Bolt, every hosted variant or other thinking settings.", "independentEvidence": [] }, { "slug": "pareto-26-9", "catalogId": "unbiased/pareto-26.9", "name": "Pareto 26.9", "provider": "Unbiased", "released": "2026-09-17", "catalogAdded": "2026-10-13", "checked": "2026-11-03", "verification": "Composite system", "note": "Pareto combines answers from several models for research, coding and multi-step work. This named September version is newly covered here; it is not an October base-model launch.", "context": 262144, "maxOutput": 131072, "modalities": [ "text", "image" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/unbiased/pareto-26.9", "officialUrl": "https://unbiased.ai/how/", "rates": { "input": 2.5, "output": 7.5, "cached": 0.25, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilityNote": "OpenRouter lists text and image input, tools and a 262,144-token context. The composite uses several models rather than one standalone checkpoint.", "accessNote": "Use the exact OpenRouter ID shown here to keep this September offer. On the native Unbiased API, pareto now selects stable 26.10; the October 15 announcement says to use pareto-26.9 to retain the older version. The API reference accepts unknown model names, so confirm the version that actually answers rather than relying only on a successful response.", "pricingNote": "The displayed $2.50 input, $0.25 cache read and $7.50 output rates are this named OpenRouter offer. Composite provider evaluations are not an independent matched score.", "apiDocumentationUrls": [ "https://unbiased.ai/developers", "https://unbiased.ai/how/" ], "independentEvidence": [] }, { "slug": "summarize", "catalogId": "scaledown/summarize", "name": "ScaleDown Summarize", "provider": "ScaleDown", "released": null, "catalogAdded": "2026-10-13", "checked": "2026-11-03", "verification": "Specialized service", "note": "Turn a long document into a shorter summary, with a focus or format you specify. Check that important details survive before using it in another workflow.", "context": 1000000, "maxOutput": 4000, "modalities": [ "text", "image", "file" ], "tools": false, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/scaledown/summarize", "officialUrl": "https://docs.scaledown.ai/api-reference/summarize-overview", "rates": { "input": 0.04, "output": 0.0, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "providerId": "summarize", "specialized": "summarization", "discovered": "2026-10-14", "capabilityNote": "The native API reads text or extracts text from supported images and PDFs. Set a focus or format in the system message. Check the summary against the document; the provider\u2019s faithfulness claim is not an independent guarantee.", "contextNote": "The selected OpenRouter offer lists 1,000,000 context tokens. These are hosted-service limits, not a verified native API allowance.", "pricingNote": "OpenRouter lists $0.04 per million input tokens and no output charge. Caching and other fees are unverified. This specialized service has no comparable general-chat quality score.", "apiDocumentationUrls": [ "https://docs.scaledown.ai/api-reference/summarize" ], "accessNote": "Use the exact specialized request format. A normal chat prompt may not supply the fields or document format the service needs.", "maxOutputNote": "Native Summarize caps answers at 1,000 tokens; larger requested values are capped. OpenRouter lists 4,000, so confirm the actual limit on your chosen service. Check the returned finish reason for a cut-off summary.", "independentEvidence": [] }, { "slug": "extract", "catalogId": "scaledown/extract", "name": "ScaleDown Extract", "provider": "ScaleDown", "released": null, "catalogAdded": "2026-10-13", "checked": "2026-11-03", "verification": "Specialized service", "note": "Pull fields such as invoice numbers, dates and customer names from documents. Describe the fields you want before sending the text.", "context": 1000000, "maxOutput": 900000, "modalities": [ "text", "image", "file" ], "tools": false, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/scaledown/extract", "officialUrl": "https://docs.scaledown.ai/api-reference/extract-overview", "rates": { "input": 0.04, "output": 0.0, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "providerId": "extract", "specialized": "extraction", "discovered": "2026-10-14", "capabilityNote": "The native API reads text or extracts text from supported images and PDFs. Describe fields with a JSON schema or entity descriptions; each match includes a confidence score and surrounding context. The native API returns at most 30 distinct values per entity type; repeated occurrences of one value are grouped. Split work when you need more matches.", "contextNote": "The selected OpenRouter offer lists 1,000,000 context tokens. These are hosted-service limits, not a verified native API allowance.", "pricingNote": "OpenRouter lists $0.04 per million input tokens and no output charge. Caching and other fees are unverified. This specialized service has no comparable general-chat quality score.", "apiDocumentationUrls": [ "https://docs.scaledown.ai/api-reference/extract" ], "accessNote": "Use the exact specialized request format. A normal chat prompt may not supply the fields or document format the service needs.", "independentEvidence": [] }, { "slug": "lfm-2-5-24b-a2b", "catalogId": "liquid/lfm-2.5-24b-a2b", "name": "LFM2.5-24B-A2B", "provider": "Liquid AI", "released": "2026-10-13", "catalogAdded": "2026-10-13", "checked": "2026-11-03", "verification": "Provider verified", "note": "A text model for assistants that reason and use tools. You can use a hosted service or download the files to run it on your own hardware.", "context": 128000, "maxOutput": 8192, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/liquid/lfm-2.5-24b-a2b", "officialUrl": "https://www.liquid.ai/blog/lfm2-5-24b-a2b", "rates": { "input": 0.05, "output": 0.7, "cached": 0.025, "write": null, "overrides": [] }, "minimumCacheTokens": null, "providerId": "LiquidAI/LFM2.5-24B-A2B", "capabilityNote": "This 24B model activates about 2.45B parameters per token and always thinks before answering. Liquid recommends tools or retrieval for knowledge-heavy work. Your application runs tool calls and checks the results.", "contextNote": "The downloadable model documents 131,072 context tokens. This OpenRouter offer has a 128,000-token context and 8,192-token output cap. Hosted limits are not interchangeable with the download.", "accessNote": "Local use needs suitable hardware and setup. The Liquid license restricts commercial use above its $10 million annual-revenue threshold; check the license for your organization. GGUF, FP8, ONNX and MLX files are formats of this model, not separate quality results.", "pricingNote": "Displayed prices apply to the OpenRouter service, not to running the download. Liquid\u2019s published speed and task scores are vendor evidence; comparable independent results remain pending.", "apiDocumentationUrls": [ "https://huggingface.co/LiquidAI/LFM2.5-24B-A2B", "https://huggingface.co/LiquidAI/LFM2.5-24B-A2B/blob/main/LICENSE" ], "independentEvidence": [] }, { "slug": "gemini-4-argon", "catalogId": "google/gemini-4-argon", "name": "Gemini 4 Argon", "provider": "Google", "released": "2026-10-13", "catalogAdded": "2026-10-13", "checked": "2026-11-03", "verification": "Public API available", "note": "Use Argon for demanding coding, research and tasks with several steps. Public API access began October 13. Choose low, medium or high thinking; medium is the default.", "context": 1048576, "maxOutput": 65536, "modalities": [ "audio", "file", "image", "text", "video" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/google/gemini-4-argon", "officialUrl": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/", "rates": { "input": 2.0, "output": 10.0, "cached": 0.1, "write": 0.1875, "overrides": [] }, "minimumCacheTokens": null, "providerId": "gemini-4-argon", "announced": "2026-09-30", "discovered": "2026-10-03", "supportedEfforts": [ "low", "medium", "high" ], "defaultEffort": "medium", "capabilityNote": "Reads text, images, video, audio and PDFs and returns text. It supports tools, structured answers and search; your application supplies permissions and runs its own tools.", "contextNote": "Input, thinking and answer share 1,048,576 tokens. Standard synchronous Interactions API output is capped at 65,536. Longer answers need background execution or continuation; the advertised 1M is not a single ordinary request allowance. OpenRouter now declares a larger 943,718 output allowance; the calculator keeps the native standard 65,536 cap until equivalent long-output service behavior and billing are verified.", "maxOutputNote": "Long decode continues in steps of up to 64K tokens. Every step rereads earlier input and output, so billed input can exceed the final document size. Repeated input is eligible for implicit caching; use the response usage totals to estimate cost.", "pricingNote": "Standard launch prices through November 30, 2026 are $2 input, $0.10 cached input and $10 output per million tokens. From December 1 they become $4/$0.20/$20. Thinking is billed as output. Batch and Flex have separate half-price rates; Priority costs more. The calculator uses the standard routed offer and excludes continuation overhead and search fees. An invite-only Contributor tier offers lower prices in exchange for allowing Google to improve its products with the data. Those are not the standard prices shown here. Google counts every processed token, including thinking and the tool instructions and function definitions added for grounded requests.", "accessNote": "Available through the paid Gemini API; no Free tier or Google AI Pro access. Caching is automatic only: explicit cache storage is not supported. The specialized cyber version still requires Fairwind approval; ordinary Argon access does not unlock it.", "apiDocumentationUrls": [ "https://ai.google.dev/gemini-api/docs/models/gemini-4-argon", "https://ai.google.dev/gemini-api/docs/latest-model", "https://ai.google.dev/gemini-api/docs/long-decode", "https://ai.google.dev/gemini-api/docs/pricing" ], "independentEvidence": [ { "source": "Arena text preference", "modelId": "gemini-4-argon-high", "score": 1525, "uncertainty": 9, "votes": 4892, "preliminary": true, "sourceDate": "2026-10-08", "checked": "2026-10-13", "sourceUrl": "https://arena.ai/leaderboard/text", "note": "Preference evidence for this exact configuration. It is not a percent-correct score or an effort sweep. This row is not merged with a differently configured LiveBench result." } ] }, { "slug": "muse-wonder", "catalogId": "meta/muse-wonder", "name": "Muse Wonder", "provider": "Meta", "released": "2026-10-12", "catalogAdded": "2026-10-12", "checked": "2026-11-03", "verification": "Provider verified", "note": "For complex coding and tasks that need several steps. One independent LiveBench result tests max thinking; it does not tell us whether lower settings are faster or better value.", "context": 1048576, "maxOutput": 943718, "modalities": [ "text", "image", "video", "file", "audio" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "muse-wonder", "sourceUrl": "https://openrouter.ai/meta/muse-wonder", "officialUrl": "https://dev.meta.ai/docs/models", "rates": { "input": 4.0, "output": 20.0, "cached": 0.2, "write": null, "overrides": [] }, "minimumCacheTokens": null, "providerId": "muse-wonder-1.0", "apiDocumentationUrls": [ "https://dev.meta.ai/docs/pricing-rate-limits", "https://dev.meta.ai/docs/reasoning" ], "capabilityNote": "Accepts text, images, video, audio and PDFs and returns text. Your application supplies tools and permissions. The native API offers minimal, low, medium, high, xhigh and max thinking; none is unsupported.", "contextNote": "The native context window is 1,048,576 tokens. The selected OpenRouter offer lists 943,718 maximum output tokens; Gateway lists different output and effort options. Reserve space for input and check the exact service.", "pricingNote": "Native Standard rates are $4 input, $0.20 cached input and $20 output per million tokens, matching this routed quote. Reasoning tokens count as billed output. Web search adds $2.50 per 1,000 queries. There is no native long-context premium.", "accessNote": "The API uses Standard mode, which Meta says excludes content from model training. Muse Code subscriptions default to Contributor mode; turn it off if you do not want submitted content used for training. Max effort on Wonder does not select Wonder Full. Meta currently excludes South Korea from Wonder access through both Model API and Muse Code.", "independentEvidence": [] }, { "slug": "ling-3-0-flash-sante", "catalogId": "inclusionai/ling-3.0-flash-sante", "name": "Ling 3.0 Flash Sante", "provider": "InclusionAI", "released": null, "catalogAdded": "2026-09-04", "checked": "2026-11-03", "verification": "Offer missing from current catalog", "note": "This health-focused text model\u2019s OpenRouter offer is no longer listed. Current access is unverified, and its earlier domain claims do not establish clinical reliability.", "context": 262144, "maxOutput": 32768, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/inclusionai/ling-3.0-flash-sante", "officialUrl": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [], "availability": "unverified", "accessNote": "Check an available endpoint before relying on this offer. Absence from one catalog does not establish that the underlying model has been retired everywhere.", "pricingNote": "Current access and prices are unverified. Earlier quotes remain in the dated downloads; they are not a current offer.", "contextNote": "The capacity shown above is from the last listed offer, not a verified current service limit.", "capabilityNote": "The last available catalog described text input and tool support. Current access is unverified." }, { "slug": "step-5-preview", "catalogId": "stepfun/step-5-preview", "name": "Step 5 Preview", "provider": "StepFun", "released": null, "catalogAdded": "2026-10-08", "checked": "2026-11-03", "verification": "Provider documented preview", "note": "A new routed offer for StepFun\u2019s coding and knowledge-work model. Native documentation supports text, image and video input, tools, 1M context and 64K output. The Gateway output allowance differs; check your exact endpoint. No matched independent score is assigned.", "context": 1048576, "maxOutput": 64000, "modalities": [ "text", "image", "video" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/stepfun/step-5-preview", "officialUrl": "https://platform.stepfun.com/docs/zh/guides/models/step-5-preview", "rates": { "input": 1.0, "output": 2.7, "cached": 0.1, "write": null, "overrides": [] }, "minimumCacheTokens": null, "supportedEfforts": [ "low", "medium", "high" ], "defaultEffort": null, "discovered": "2026-10-09", "dateNote": "OpenRouter addition is October 8. Gateway has an older release-metadata date; neither is treated as a verified native launch date.", "contextNote": "Native and OpenRouter documentation agree on 1M context and 64K output. Gateway currently lists 1M max_tokens and text/image input; that serving declaration does not override the native 64K cap or establish video support there.", "capabilityNote": "Native API ID step-5-preview accepts text, images and video; supports tools, JSON/JSON Schema and low/medium/high effort. The application provides tools and external access. The routed offer defaults to medium, but the native default is not explicitly established by the reviewed reference.", "pricingNote": "Native StepFun billing is in yuan per million tokens: \u00a57 for uncached input and \u00a520 for output. At 00:30 Beijing time on October 14 (October 13 at 16:30 UTC), the standard cache-read rate increased from \u00a50.35 to \u00a50.70. Users who called Step 5 Preview through the pay-as-you-go API before that cutoff keep \u00a50.35 for later calls. Step Plan access is excluded from this change. The dollar prices shown here are separate OpenRouter quotes; they do not establish your native account rate.", "apiDocumentationUrls": [ "https://platform.stepfun.com/docs/zh/announcements/step-5-preview-price-adjustment" ], "independentEvidence": [] }, { "slug": "claude-haiku-5-5", "catalogId": "anthropic/claude-haiku-5.5", "name": "Claude Haiku 5.5", "provider": "Anthropic", "released": "2026-10-07", "catalogAdded": "2026-10-07", "checked": "2026-11-03", "verification": "Provider verified", "note": "Five native effort levels with medium default; thinking is on by default. The 1M context and 128K output limits are documented. Prompts above 100K use five times the base token rates. Its newer tokenizer can count the same text differently from Haiku 4.5; compare actual billed usage before switching.", "context": 1000000, "maxOutput": 128000, "modalities": [ "text", "image", "file" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "claude-haiku-5-5", "sourceUrl": "https://openrouter.ai/anthropic/claude-haiku-5.5", "officialUrl": "https://platform.claude.com/docs/en/models/haiku-5-5/overview", "rates": { "input": 0.1, "output": 0.5, "cached": 0.01, "write": 0.125, "overrides": [ { "min_prompt_tokens": 100000, "prompt": 0.5, "completion": 2.5, "input_cache_read": 0.05, "input_cache_write": 0.625, "input_cache_write_1h": "0.000001" } ] }, "minimumCacheTokens": 512, "providerId": "claude-haiku-5-5", "supportedEfforts": [ "low", "medium", "high", "xhigh", "max" ], "defaultEffort": "medium", "thinkingDefault": "adaptive", "capabilityNote": "Thinking can be disabled at low, medium or high effort, but that request is rejected at xhigh/max. Manual budget_tokens thinking is unsupported. Omit sampling parameters and handle thinking blocks by type. The newer tokenizer counts about 30% more tokens for the same text on average; actual usage varies.", "priceContract": { "threshold": 100000, "comparison": "strictly greater", "base": { "input": 0.1, "output": 0.5, "cached": 0.01, "write": 0.125, "write1h": 0.2 }, "long": { "input": 0.5, "output": 2.5, "cached": 0.05, "write": 0.625, "write1h": 1 }, "sourceUrl": "https://platform.claude.com/docs/en/about-claude/pricing" }, "independentEvidence": [] }, { "slug": "mistral-large-4-0", "catalogId": "mistralai/mistral-large-4-0", "name": "Mistral Large 4", "provider": "Mistral", "released": "2026-10-06", "catalogAdded": "2026-10-06", "checked": "2026-11-03", "verification": "Public preview", "note": "Preview API access is available; open weights remain planned for later October. The native model page no longer shows its launch discount and lists $1.36 input, $0.14 cached input and $4.18 output per million. This routed offer has its own current prices.", "context": 1048576, "maxOutput": 262144, "modalities": [ "text", "image" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "mistral-large-4-0", "sourceUrl": "https://openrouter.ai/mistralai/mistral-large-4-0", "officialUrl": "https://docs.mistral.ai/models/mistral-large-4-0", "rates": { "input": 0.7, "output": 2.1, "cached": 0.07, "write": null, "overrides": [] }, "minimumCacheTokens": null, "accessNote": "Official FP8 and NVFP4 weights are downloadable as of October 29. They are formats of the same model. The Mistral Ecosystem Open-Weight License has extra commercial licensing conditions for model-as-a-service businesses; read it before offering hosted access.", "apiDocumentationUrls": [ "https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B", "https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B", "https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B/blob/main/LICENSE" ], "independentEvidence": [] }, { "slug": "switchyard", "catalogId": "nvidia/switchyard", "name": "NVIDIA Switchyard", "provider": "NVIDIA", "released": null, "catalogAdded": "2026-09-21", "checked": "2026-11-03", "verification": "Routing system", "note": "A router over a configured model pool, not a new base model. OpenRouter now explicitly states that each request uses its serving model\u2019s price and context allowance. Negative catalog price sentinels remain unknown fixed rates; no free service or guaranteed one-million-token request is inferred.", "context": 1000000, "maxOutput": null, "modalities": [ "text" ], "tools": false, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/nvidia/switchyard", "officialUrl": "https://github.com/NVIDIA-NeMo/Switchyard", "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [] }, { "slug": "pareto-26-10-preview", "catalogId": "unbiased/pareto-26.10-preview", "name": "Pareto 26.10 Preview", "provider": "Unbiased", "released": "2026-10-01", "catalogAdded": "2026-10-01", "checked": "2026-11-03", "verification": "Offer missing from current catalog", "note": "The OpenRouter preview offer is no longer listed. Unbiased has released stable Pareto 26.10 through its own service.", "context": 1048576, "maxOutput": 131072, "modalities": [ "text", "image" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/unbiased/pareto-26.10-preview", "officialUrl": "https://unbiased.ai/blog/pareto-26-10-preview/", "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [], "availability": "unverified", "pricingNote": "Current prices for this old preview route are unknown. Historical quotes remain in dated downloads.", "contextNote": "Capacity shown is from the last listed preview offer.", "accessNote": "For the native stable release, use pareto-26.10. An absent OpenRouter preview does not mean the native model has shut down." }, { "slug": "apodex-1-1-mini-free", "catalogId": "apodex/apodex-1.1-mini:free", "name": "Apodex 1.1 Mini (free route)", "provider": "Apodex", "released": null, "catalogAdded": "2026-10-01", "checked": "2026-11-03", "verification": "Offer missing from current catalog", "note": "This reasoning and tool-use model\u2019s free OpenRouter offer is no longer listed. Confirm another available offer before connecting an application.", "context": 262144, "maxOutput": 235929, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/apodex/apodex-1.1-mini:free", "officialUrl": "https://platform.apodex.ai/docs/models", "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [], "availability": "unverified", "accessNote": "Check an available endpoint before relying on this offer. Absence from one catalog does not establish that the underlying model has been retired everywhere.", "pricingNote": "Current access and prices are unverified. Earlier quotes remain in the dated downloads; they are not a current offer.", "contextNote": "The capacity shown above is from the last listed offer, not a verified current service limit.", "capabilityNote": "The last available catalog described text input and tool support. Current access is unverified." }, { "slug": "gpt-6-1-sol", "catalogId": "openai/gpt-6.1-sol", "name": "GPT-6.1 Sol", "provider": "OpenAI", "released": "2026-09-29", "catalogAdded": "2026-09-29", "checked": "2026-11-03", "verification": "Provider verified", "note": "Five supported effort levels: low, medium (the API default), high, extra high and max. None and minimal are unsupported. Tool calling uses Responses; Chat Completions supports requests without tools.", "context": 1050000, "maxOutput": 128000, "modalities": [ "file", "image", "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "gpt-6-1-sol", "sourceUrl": "https://openrouter.ai/openai/gpt-6.1-sol", "officialUrl": "https://developers.openai.com/api/docs/models/gpt-6.1-sol", "rates": { "input": 2.0, "output": 10.0, "cached": 0.1, "write": 2.5, "overrides": [ { "min_prompt_tokens": 272000, "prompt": 4.0, "completion": 15.0, "input_cache_read": 0.2, "input_cache_write": 5.0 } ] }, "minimumCacheTokens": null, "independentEvidence": [] }, { "slug": "claude-sonnet-5-5", "catalogId": "anthropic/claude-sonnet-5.5", "name": "Claude Sonnet 5.5", "provider": "Anthropic", "released": "2026-09-28", "catalogAdded": "2026-09-28", "checked": "2026-11-03", "verification": "Provider verified", "note": "Adaptive thinking, high default effort and 128K maximum output. Migration changes affect forced tool use, thinking replay and older computer-use tools.", "context": 1000000, "maxOutput": 128000, "modalities": [ "text", "image", "file" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "claude-sonnet-5-5", "sourceUrl": "https://openrouter.ai/anthropic/claude-sonnet-5.5", "officialUrl": "https://platform.claude.com/docs/en/models/sonnet-5-5/overview", "rates": { "input": 2.0, "output": 10.0, "cached": 0.1, "write": 2.5, "overrides": [] }, "minimumCacheTokens": 512, "independentEvidence": [ { "source": "Arena text preference", "modelId": "claude-sonnet-5.5-xhigh", "score": 1476, "uncertainty": 9, "votes": 4838, "preliminary": false, "sourceDate": "2026-10-08", "checked": "2026-10-13", "sourceUrl": "https://arena.ai/leaderboard/text", "note": "Preference evidence for this exact configuration. It is not a percent-correct score or an effort sweep. This row is not merged with a differently configured LiveBench result." } ] }, { "slug": "gpt-6-1-sol-pro", "catalogId": "openai/gpt-6.1-sol-pro", "name": "GPT-6.1 Sol Pro", "provider": "OpenAI", "released": null, "catalogAdded": "2026-09-29", "checked": "2026-11-03", "verification": "Catalog variant", "note": "OpenRouter describes this as GPT-6.1 Sol served in pro mode, with greater token use. It is a serving variant; no independent result is assigned from the base model.", "context": 1050000, "maxOutput": 128000, "modalities": [ "file", "image", "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/openai/gpt-6.1-sol-pro", "officialUrl": null, "rates": { "input": 2.0, "output": 10.0, "cached": 0.1, "write": 2.5, "overrides": [ { "min_prompt_tokens": 272000, "prompt": 4.0, "completion": 15.0, "input_cache_read": 0.2, "input_cache_write": 5.0 } ] }, "minimumCacheTokens": null, "independentEvidence": [] }, { "slug": "glm-5-3-prime", "catalogId": "z-ai/glm-5.3-prime", "name": "GLM 5.3 Prime", "provider": "Z.ai", "released": null, "catalogAdded": "2026-09-23", "checked": "2026-11-03", "verification": "Catalog variant", "note": "Separate catalog endpoint. A throughput claim does not establish measured speed here; base-model benchmark scores are withheld.", "context": 1000000, "maxOutput": 131072, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/z-ai/glm-5.3-prime", "officialUrl": null, "rates": { "input": 2.8, "output": 8.8, "cached": 0.56, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [] }, { "slug": "qwen3-8-max-prime", "catalogId": "qwen/qwen3.8-max-prime", "name": "Qwen3.8 Max Prime", "provider": "Alibaba", "released": null, "catalogAdded": "2026-09-23", "checked": "2026-11-03", "verification": "Catalog variant", "note": "Separate catalog endpoint. No exact matched benchmark result or measured speed is verified for this variant.", "context": 1000000, "maxOutput": 131072, "modalities": [ "text", "image", "video" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/qwen/qwen3.8-max-prime", "officialUrl": null, "rates": { "input": 4.0, "output": 12.0, "cached": 0.5, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [] }, { "slug": "ember-1", "catalogId": "fireworks/ember-1", "name": "Ember-1", "provider": "Fireworks", "released": "2026-09-23", "catalogAdded": "2026-09-24", "checked": "2026-11-03", "verification": "Provider verified", "note": "Research preview announced September 23 with temporary two-week serverless access. Vendor claims about shorter reasoning are not independently measured task costs; verify availability.", "context": 1048576, "maxOutput": 943718, "modalities": [ "text", "image" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/fireworks/ember-1", "officialUrl": "https://fireworks.ai/blog/ember-1", "rates": { "input": 3.0, "output": 15.0, "cached": 0.3, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [] }, { "slug": "aion-3-5", "catalogId": "aion-labs/aion-3.5", "name": "Aion 3.5", "provider": "AionLabs", "released": "2026-09-23", "catalogAdded": "2026-09-23", "checked": "2026-11-03", "verification": "Provider documented composite", "note": "Catalog-listed text generation system. Independent general-task or storytelling performance has not been verified in our matched datasets.", "context": 262144, "maxOutput": 32768, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/aion-labs/aion-3.5", "officialUrl": null, "rates": { "input": 3.0, "output": 6.0, "cached": 0.75, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilitySourceUrl": "https://www.aionlabs.ai/docs/models/", "apiDocumentationUrls": [ "https://api.aionlabs.ai/docs/api-reference/" ], "capabilityNote": "AionLabs describes a GLM-family system that uses several cooperating models. It is a composite system; individual constituent scores are not assigned to this endpoint. Native documentation verifies the September 23 release and 256K input context with up to 32K output. The selected routed offer retains its separately observed limits and price.", "independentEvidence": [] }, { "slug": "aion-3-5-mini", "catalogId": "aion-labs/aion-3.5-mini", "name": "Aion 3.5 Mini", "provider": "AionLabs", "released": "2026-09-23", "catalogAdded": "2026-09-23", "checked": "2026-11-03", "verification": "Provider documented composite", "note": "Lower-priced text generation endpoint. A lower price does not establish equal quality or cost per successful task.", "context": 262144, "maxOutput": 32768, "modalities": [ "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/aion-labs/aion-3.5-mini", "officialUrl": null, "rates": { "input": 0.7, "output": 1.4, "cached": 0.18, "write": null, "overrides": [] }, "minimumCacheTokens": null, "capabilitySourceUrl": "https://www.aionlabs.ai/docs/models/", "apiDocumentationUrls": [ "https://api.aionlabs.ai/docs/api-reference/" ], "capabilityNote": "AionLabs describes a GLM-family system that uses several cooperating models. It is a composite system; individual constituent scores are not assigned to this endpoint. Native documentation verifies the September 23 release and 256K input context with up to 32K output. The selected routed offer retains its separately observed limits and price.", "independentEvidence": [] }, { "slug": "perceptron-mk1-5", "catalogId": "perceptron/perceptron-mk1.5", "name": "Perceptron Mk1.5", "provider": "Perceptron", "released": "2026-09-25", "catalogAdded": "2026-09-25", "checked": "2026-11-03", "verification": "Provider verified", "note": "Provider verifies text, image, audio and video input and 32K context; this catalog offer advertises 36,864 tokens. Confirm the selected endpoint limit. Specialized performance is not scored using another model\u2019s text benchmarks.", "context": 36864, "maxOutput": 8192, "modalities": [ "text", "image", "video", "audio" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/perceptron/perceptron-mk1.5", "officialUrl": "https://www.perceptron.inc/blog/introducing-perceptron-mk1-5", "rates": { "input": 0.15, "output": 1.5, "cached": null, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [] } ], "baselines": [ { "slug": "gpt-6-sol", "catalogId": "openai/gpt-6-sol", "name": "GPT-6 Sol", "provider": "OpenAI", "released": null, "catalogAdded": "2026-09-22", "checked": "2026-11-03", "verification": "Catalog verified", "note": "", "context": 1050000, "maxOutput": 128000, "modalities": [ "file", "image", "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "gpt-6-sol", "sourceUrl": "https://openrouter.ai/openai/gpt-6-sol", "officialUrl": null, "rates": { "input": 2.0, "output": 10.0, "cached": 0.2, "write": 2.5, "overrides": [ { "min_prompt_tokens": 272000, "prompt": 4.0, "completion": 15.0, "input_cache_read": 0.4, "input_cache_write": 5.0 } ] }, "minimumCacheTokens": null }, { "slug": "claude-haiku-4-5", "catalogId": "anthropic/claude-haiku-4.5", "name": "Claude Haiku 4.5", "provider": "Anthropic", "released": null, "catalogAdded": "2025-10-15", "checked": "2026-11-03", "verification": "Catalog verified", "note": "", "context": 200000, "maxOutput": 64000, "modalities": [ "text", "image", "file" ], "tools": true, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "sourceUrl": "https://openrouter.ai/anthropic/claude-haiku-4.5", "officialUrl": null, "rates": { "input": 1.0, "output": 5.0, "cached": 0.1, "write": 1.25, "overrides": [] }, "minimumCacheTokens": 4096 }, { "slug": "gpt-6-astra", "catalogId": "openai/gpt-6-astra", "name": "GPT-6 Astra", "provider": "OpenAI", "released": null, "catalogAdded": "2026-09-04", "checked": "2026-11-03", "verification": "Catalog verified", "note": "", "context": 1050000, "maxOutput": 128000, "modalities": [ "file", "image", "text" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "gpt-6-astra", "sourceUrl": "https://openrouter.ai/openai/gpt-6-astra", "officialUrl": null, "rates": { "input": 10.0, "output": 50.0, "cached": 1.0, "write": 12.5, "overrides": [ { "min_prompt_tokens": 272000, "prompt": 20.0, "completion": 75.0, "input_cache_read": 2.0, "input_cache_write": 25.0 } ] }, "minimumCacheTokens": null }, { "slug": "claude-sonnet-5", "catalogId": "anthropic/claude-sonnet-5", "name": "Claude Sonnet 5", "provider": "Anthropic", "released": null, "catalogAdded": "2026-06-30", "checked": "2026-11-03", "verification": "Catalog verified", "note": "", "context": 1000000, "maxOutput": 128000, "modalities": [ "text", "image", "file" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "claude-sonnet-5", "sourceUrl": "https://openrouter.ai/anthropic/claude-sonnet-5", "officialUrl": null, "rates": { "input": 2.0, "output": 10.0, "cached": 0.2, "write": 2.5, "overrides": [] }, "minimumCacheTokens": null }, { "slug": "claude-opus-5-5", "catalogId": "anthropic/claude-opus-5.5", "name": "Claude Opus 5.5", "provider": "Anthropic", "released": null, "catalogAdded": "2026-09-22", "checked": "2026-11-03", "verification": "Catalog verified", "note": "", "context": 1000000, "maxOutput": 128000, "modalities": [ "text", "image", "file" ], "tools": true, "benchmarkStatus": "Matched LiveBench", "benchmarkSlug": "claude-opus-5-5", "sourceUrl": "https://openrouter.ai/anthropic/claude-opus-5.5", "officialUrl": null, "rates": { "input": 4.0, "output": 20.0, "cached": 0.2, "write": 5.0, "overrides": [] }, "minimumCacheTokens": null } ], "checks": [ { "source": "LiveBench", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://livebench.ai/table_2026_06_25.csv", "checked": "2026-11-03" }, { "source": "Arena", "status": "403; retained October 13 retrieval and October 8 source date.", "url": "https://arena.ai/leaderboard/text", "checked": "2026-11-03" }, { "source": "Hex DataBench", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://hex.tech/databench/", "checked": "2026-11-03" }, { "source": "OckBench", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://ockbench.github.io/static/data/top200_model_performance.csv", "checked": "2026-11-03" }, { "source": "OpenRouter", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://openrouter.ai/api/v1/models", "checked": "2026-11-03" }, { "source": "OpenAI", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://developers.openai.com/api/docs/changelog", "checked": "2026-11-03" }, { "source": "Anthropic", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://platform.claude.com/docs/en/release-notes/overview", "checked": "2026-11-03" }, { "source": "Google", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://ai.google.dev/gemini-api/docs/changelog", "checked": "2026-11-03" }, { "source": "DeepSeek", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://api-docs.deepseek.com/quick_start/pricing/", "checked": "2026-11-03" }, { "source": "xAI", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://docs.x.ai/developers/release-notes", "checked": "2026-11-03" }, { "source": "Alibaba/Qwen", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://huggingface.co/Qwen", "checked": "2026-11-03" }, { "source": "Z.ai", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://docs.z.ai/release-notes/new-released", "checked": "2026-11-03" }, { "source": "Moonshot", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://www.moonshot.ai/", "checked": "2026-11-03" }, { "source": "MiniMax", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://agent.minimax.io/tools/m3-1-flash-preview", "checked": "2026-11-03" }, { "source": "Meta", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://dev.meta.ai/docs/models", "checked": "2026-11-03" }, { "source": "NVIDIA", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://developer.nvidia.com/topics/ai/nemotron", "checked": "2026-11-03" }, { "source": "Thinking Machines", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://tinker-docs.thinkingmachines.ai/tinker/models/", "checked": "2026-11-03" }, { "source": "BFCL", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://gorilla.cs.berkeley.edu/data_overall.csv", "checked": "2026-11-03" }, { "source": "Vectara", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://github.com/vectara/hallucination-leaderboard", "checked": "2026-11-03" }, { "source": "Fireworks", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://fireworks.ai/blog/ember-1", "checked": "2026-11-03" }, { "source": "Perceptron", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://www.perceptron.inc/blog/introducing-perceptron-mk1-5", "checked": "2026-11-03" }, { "source": "InclusionAI / Vercel", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://vercel.com/changelog/ling-3-1-flash-is-now-available-on-ai-gateway", "checked": "2026-11-03" }, { "source": "Apodex", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://platform.apodex.ai/docs/models", "checked": "2026-11-03" }, { "source": "Unbiased", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://unbiased.ai/blog/pareto-26-10-preview/", "checked": "2026-11-03" }, { "source": "Perplexity", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://docs.perplexity.ai/docs/decisions/quickstart", "checked": "2026-11-03" }, { "source": "Cloudflare", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://blog.cloudflare.com/clef-decision-models/", "checked": "2026-11-03" }, { "source": "Historical matched effort studies", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://www.stet.sh/blog/gpt-55-codex-graphql-reasoning-curve", "checked": "2026-11-03" }, { "source": "Vercel AI Gateway exact model catalog", "url": "https://ai-gateway.vercel.sh/v1/models", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Convai Innovations / Vercel", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://vercel.com/changelog/laya-decision-model-now-available-on-ai-gateway-free-through-october-31", "checked": "2026-11-03" }, { "source": "Artificial Analysis", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://artificialanalysis.ai/terms-of-use", "checked": "2026-11-03" }, { "source": "Mistral", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://docs.mistral.ai/resources/changelogs", "checked": "2026-11-03" }, { "source": "TypeSafe AI / Vercel", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://vercel.com/changelog/typesafe-ai-jev-now-available-on-ai-gateway", "checked": "2026-11-03" }, { "source": "Vercel production usage", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://vercel.com/docs/ai-gateway/leaderboards", "checked": "2026-11-03" }, { "source": "Reflection", "url": "https://developers.reflection.ai/models", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Reka", "url": "https://reka.ai/news/rho-1-collapsing-the-multimodal-stack", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Aleph Alpha", "url": "https://huggingface.co/Aleph-Alpha/Kolibri-1", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Ai2", "url": "https://allenai.org/blog/astabrief", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Hex DataBench reuse permissions", "url": "https://learn.hex.tech/docs/legal/terms-and-conditions-2020-03-27", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "StepFun", "url": "https://platform.stepfun.com/docs/zh/guides/models/step-5-preview", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Callstack", "url": "https://apex.callstack.com/models", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Liquid AI", "url": "https://www.liquid.ai/news/models", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Ai2 announcements", "url": "https://allenai.org/news", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Microsoft", "url": "https://commandline.microsoft.com/microsoft-decision-1-model-foundry/", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "FINAL-Bench", "url": "https://huggingface.co/FINAL-Bench/Darwin-27B-ZTC-v2", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "LightOn", "url": "https://huggingface.co/blog/lightonai/lightonocr-3", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "TII / Falcon", "url": "https://huggingface.co/blog/tiiuae/falcon-emirati", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Hugging Face discovery", "url": "https://huggingface.co/blog", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Fireworks FireRouter", "url": "https://fireworks.ai/blog/introducing-firerouter-with-opus", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "AionLabs native models", "url": "https://www.aionlabs.ai/docs/models/", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Additional provider feeds", "url": "https://fireworks.ai/blog", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Nace.AI", "url": "https://console.nace.ai/docs/drex/reference/models", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Cloudflare native decision contracts", "url": "https://developers.cloudflare.com/workers-ai/models/", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "JetBrains Mellum", "url": "https://blog.jetbrains.com/ai/category/releases/", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "ScaleDown", "url": "https://docs.scaledown.ai", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "FrontiersMind", "url": "https://huggingface.co/blog/FrontiersMind/lumma-fev-27b", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "DMIS Lab", "url": "https://huggingface.co/blog/dmis-lab/ultrabert", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Moondream", "url": "https://moondream.ai/blog/moondream-3-1-beyond-benchmarks", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Swiss AI", "url": "https://huggingface.co/blog/swiss-ai/apertus-v15-release-notes", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "TypeSafe AI", "url": "https://docs.typesafe.ai/models", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Daily / Cloudflare", "url": "https://developers.cloudflare.com/workers-ai/models/phonellm-mini-alpha-2/", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Elorian", "url": "https://elorian.ai/blog/introducing-glimpse", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "China Telecom / TeleAI", "url": "https://huggingface.co/XingChen-AGI/Xing4.0-29B-A4B", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "IBM Granite Switch", "url": "https://huggingface.co/blog/ibm-granite/granite-switch-4-1", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Microsoft MAI research", "url": "https://microsoft.ai/news/frontier-pre-training-at-microsoft-ai/", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "checked": "2026-11-03" }, { "source": "Inception", "url": "https://docs.inceptionlabs.ai/get-started/models", "checked": "2026-11-03", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates." }, { "source": "Cloudflare Clairic", "url": "https://developers.cloudflare.com/workers-ai/models/clairic/", "checked": "2026-11-03", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates." }, { "source": "Sqwish", "url": "https://docs.sqwish.ai/guides/models-and-fallback", "checked": "2026-11-03", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates." }, { "source": "Jais / IFM", "url": "https://ifm.ai/blog/jais3/", "checked": "2026-11-03", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates." }, { "source": "Kimi K3.1 Pro", "url": "https://platform.kimi.ai/docs/guide/kimi-k3-1-pro-quickstart", "checked": "2026-11-03", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates." }, { "source": "DiffuSpace Clover", "checked": "2026-11-03", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates.", "url": "https://huggingface.co/blog/Diffuspace/clover-35b-a3b-preview" }, { "source": "SenseNova", "checked": "2026-11-03", "url": "https://huggingface.co/sensenova/SenseNova-Loop-9B-Preview", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates." }, { "source": "Giotto", "checked": "2026-11-03", "url": "https://www.giotto.ai/api", "status": "November 3 primary-source review. Successful, unavailable and changed responses are recorded in the dated receipt; historical results retain their actual dates." }, { "source": "LongCat", "checked": "2026-11-03", "url": "https://longcat.ai/platform/docs/ChangeLog.html", "status": "Primary public documentation reviewed. Access limitations and conflicting limits retained; no inherited benchmark scores." }, { "source": "Deep Principle", "checked": "2026-11-03", "url": "https://huggingface.co/deep-principle/Hydrogen-1.0", "status": "Primary public documentation reviewed. Access limitations and conflicting limits retained; no inherited benchmark scores." } ], "sourceHash": "39a64d8c2f8c542efd9899b923e265c10bee58a5664fb306ab95b487f69ba3d4", "priceChanges": [ { "slug": "deepseek-v4-flash-0731", "name": "DeepSeek V4 Flash 0731", "catalogId": "deepseek/deepseek-v4-flash-0731", "previousDate": "2026-11-02", "currentDate": "2026-11-03", "previous": { "input": 0.119, "output": 0.238, "cached": 0.0238 }, "current": { "input": 0.059, "output": 0.3, "cached": 0.035 }, "scheduled": false, "previousScheduleRemoved": false, "note": "This hosting offer changed price. Check the service you will use; direct-provider prices may differ." }, { "slug": "deepseek-v4-1-flash", "name": "DeepSeek V4.1 Flash", "catalogId": "deepseek/deepseek-v4.1-flash", "previousDate": "2026-11-02", "currentDate": "2026-11-03", "previous": { "input": 0.045, "output": 0.5, "cached": 0.006 }, "current": { "input": 0.045, "output": 0.5, "cached": 0.012 }, "scheduled": false, "previousScheduleRemoved": false, "note": "This hosting offer changed price. Check the service you will use; direct-provider prices may differ." } ], "lifecycle": [ { "model": "Claude Sonnet 4.5", "providerId": "claude-sonnet-4-5-20250929", "status": "Deprecated; retirement scheduled", "announced": "2026-09-30", "retirement": "2026-11-30", "replacement": "claude-sonnet-5-5", "checked": "2026-11-03", "sourceUrl": "https://platform.claude.com/docs/en/about-claude/model-deprecations", "note": "Claude API retirement notice. Partner-operated Bedrock and Google Cloud schedules can differ. Historical benchmarks remain historical; scores are not transferred to the recommended replacement." }, { "model": "GLM-4.5", "providerId": "z-ai/glm-4.5", "status": "Catalog expiration scheduled", "announced": null, "retirement": "2026-12-31", "replacement": null, "checked": "2026-11-03", "sourceType": "catalog", "sourceUrl": "https://openrouter.ai/z-ai/glm-4.5", "note": "The exact OpenRouter entry now declares this expiration date. This is endpoint-specific catalog evidence; it does not establish a direct-provider shutdown or retirement of every hosted copy. No exact cutoff time or replacement is verified." }, { "model": "GLM-4.7", "providerId": "z-ai/glm-4.7", "status": "Catalog expiration scheduled", "announced": null, "retirement": "2026-12-31", "replacement": null, "checked": "2026-11-03", "sourceType": "catalog", "sourceUrl": "https://openrouter.ai/z-ai/glm-4.7", "note": "The exact OpenRouter entry now declares this expiration date. This is endpoint-specific catalog evidence; it does not establish a direct-provider shutdown or retirement of every hosted copy. No exact cutoff time or replacement is verified." }, { "model": "Dots 3 Note Preview free route", "providerId": "dots-studio/dots-3-note-preview:free", "status": "Catalog expiration scheduled", "announced": null, "retirement": "2026-12-31", "replacement": null, "checked": "2026-11-03", "sourceType": "catalog", "sourceUrl": "https://openrouter.ai/dots-studio/dots-3-note-preview:free", "note": "The exact OpenRouter entry now declares this expiration date. This is endpoint-specific catalog evidence; it does not establish a direct-provider shutdown or retirement of every hosted copy. No exact cutoff time or replacement is verified." }, { "model": "GPT-5.3 Codex", "providerId": "gpt-5.3-codex", "status": "Deprecated; retirement scheduled", "announced": "2026-10-01", "retirement": "2027-04-01", "replacement": "gpt-6-sol", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "note": "Official OpenAI API notice gives six months of notice. Historical results remain attached to the tested model; no successor scores are transferred." }, { "model": "GPT-5.4 nano", "providerId": "gpt-5.4-nano", "status": "Deprecated; retirement scheduled", "announced": "2026-10-01", "retirement": "2027-04-01", "replacement": "gpt-6-luna", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "note": "Official OpenAI API notice gives six months of notice. Historical results remain attached to the tested model; no successor scores are transferred." }, { "model": "GPT-5.1", "providerId": "gpt-5.1", "status": "Deprecated; retirement scheduled", "announced": "2026-10-01", "retirement": "2027-04-01", "replacement": "gpt-6-sol", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "note": "Official OpenAI API notice gives six months of notice. Historical results remain attached to the tested model; no successor scores are transferred." }, { "model": "GLM5.2 hosted by Mistral", "providerId": "zai-glm-5-2", "status": "Hosted endpoint deprecated; retirement scheduled", "announced": "2026-09-29", "retirement": "2026-10-31", "replacement": "zai-glm-5-3", "checked": "2026-11-03", "sourceUrl": "https://docs.mistral.ai/resources/changelogs", "note": "This notice applies to the Mistral-hosted endpoint. It does not retire every Z.ai GLM5.2 deployment. The provider states the replacement uses the same price; no benchmark score is transferred." }, { "model": "Antigravity Agent May 2026 preview", "providerId": "antigravity-preview-05-2026", "status": "Managed agent \u00b7 earliest shutdown date reached", "announced": "2026-09-17", "retirement": "2026-10-05", "replacement": "antigravity-preview-09-2026", "checked": "2026-11-03", "sourceUrl": "https://ai.google.dev/gemini-api/docs/deprecations", "note": "Google lists October 5 as the earliest shutdown date for this managed agent, rather than a base model. We have not tested whether the endpoint is already unavailable. Local-tool integrations need the new parameter and file-edit contracts described in the September 17 release notes; changing only the agent ID may be insufficient." }, { "model": "OpenAI Chat Latest", "providerId": "chat-latest", "status": "Mutable alias \u00b7 snapshot updated", "announced": "2026-10-07", "retirement": null, "replacement": null, "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/changelog", "note": "OpenAI updated this chat snapshot on October 7. It remains a moving alias, not a newly named base model or a pinned benchmark version. The provider recommends the GPT-6 family for production APIs; no predecessor score is transferred." }, { "model": "OpenRouter Haiku Latest", "providerId": "~anthropic/claude-haiku-latest", "status": "Routed alias \u00b7 now points to Haiku 5.5", "announced": null, "retirement": null, "replacement": "anthropic/claude-haiku-5.5", "checked": "2026-11-03", "sourceType": "catalog", "sourceUrl": "https://openrouter.ai/api/v1/models", "note": "The exact catalog alias changed from Haiku 4.5 to Haiku 5.5 between the October 7 and October 8 observations. Thinking defaults, tokenization and long-context prices change with it. Pin and test a specific model before depending on this alias; native Anthropic IDs are separate contracts." }, { "model": "Deep Research December 2025 agent", "providerId": "deep-research-pro-preview-12-2025", "status": "Deprecated; shutdown scheduled", "announced": "2026-10-08", "retirement": "2026-10-23", "replacement": "deep-research-preview-04-2026 or deep-research-max-preview-04-2026", "checked": "2026-11-03", "sourceUrl": "https://ai.google.dev/gemini-api/docs/changelog", "note": "This applies to the managed Deep Research agent, not the underlying Gemini model family. Update the agent parameter in interactions.create and test speed versus comprehensiveness on your research tasks." }, { "model": "DeepSeek V4.1 Flash Fast Gateway offer", "providerId": "deepseek/deepseek-v4.1-flash-fast", "status": "Gateway offer returned October 13", "announced": null, "retirement": null, "replacement": null, "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "note": "This route returned October 13 after its October 9 absence. Current output and prices differ from the previous offer; see the current profile before using it." }, { "model": "Drex native moving and retired IDs", "providerId": "drex-latest / drex-v1.1", "status": "Native alias and retired-ID fallback", "announced": null, "retirement": null, "replacement": "drex-v2.0 / drex-v1.5", "checked": "2026-11-03", "sourceUrl": "https://console.nace.ai/docs/drex/reference/models", "note": "Since October 12, omitted model and drex-latest select v2.0. Retired drex-v1.1 still selects v1.5. Current v2.0 refuses requests above its limits; the previously documented long-request fallback has been removed." }, { "model": "Step 5 Preview native cache billing", "providerId": "step-5-preview", "status": "Native cache-price change in effect", "announced": null, "retirement": null, "replacement": null, "checked": "2026-11-03", "sourceUrl": "https://platform.stepfun.com/docs/zh/announcements/step-5-preview-price-adjustment", "note": "Since October 14 at 00:30 Beijing time, the standard native cache-read price increased from \u00a50.35 to \u00a50.70 per million tokens. Users with a pay-as-you-go call before the cutoff keep \u00a50.35. Step Plan is excluded. Uncached input \u00a57 and output \u00a520 remain unchanged. These native yuan rates do not replace routed dollar quotes.", "sourceType": "pricing", "effectiveAt": "2026-10-14T00:30:00+08:00", "effectiveLabel": "October 14, 00:30 Beijing time" }, { "model": "GPT-4o August 2024", "providerId": "gpt-4o-2024-08-06", "status": "Deprecated; retirement scheduled", "announced": "2026-10-14", "retirement": "2027-01-14", "replacement": "gpt-6-luna", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "note": "OpenAI plans to remove this API snapshot on January 14, 2027. Fine-tuned versions follow the same schedule. Test the replacement on your own work; old benchmark scores do not transfer." }, { "model": "GPT-4o November 2024", "providerId": "gpt-4o-2024-11-20", "status": "Deprecated; retirement scheduled", "announced": "2026-10-14", "retirement": "2027-01-14", "replacement": "gpt-6-luna", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "note": "OpenAI plans to remove this API snapshot on January 14, 2027. Fine-tuned versions follow the same schedule. Test the replacement on your own work; old benchmark scores do not transfer." }, { "model": "GPT-4o mini", "providerId": "gpt-4o-mini-2024-07-18", "status": "Deprecated; retirement scheduled", "announced": "2026-10-14", "retirement": "2027-01-14", "replacement": "gpt-6-luna", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "note": "OpenAI plans to remove this API snapshot on January 14, 2027. Fine-tuned versions follow the same schedule. Test the replacement on your own work; old benchmark scores do not transfer." }, { "model": "GPT-4.1", "providerId": "gpt-4.1-2025-04-14", "status": "Deprecated; retirement scheduled", "announced": "2026-10-14", "retirement": "2027-01-14", "replacement": "gpt-6-luna", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "note": "OpenAI plans to remove this API snapshot on January 14, 2027. Fine-tuned versions follow the same schedule. Test the replacement on your own work; old benchmark scores do not transfer." }, { "model": "GPT-4.1 mini", "providerId": "gpt-4.1-mini-2025-04-14", "status": "Deprecated; retirement scheduled", "announced": "2026-10-14", "retirement": "2027-01-14", "replacement": "gpt-6-luna", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "note": "OpenAI plans to remove this API snapshot on January 14, 2027. Fine-tuned versions follow the same schedule. Test the replacement on your own work; old benchmark scores do not transfer." }, { "model": "Pareto 26.10 Preview on OpenRouter", "providerId": "unbiased/pareto-26.10-preview", "status": "Preview route absent; native stable release available", "announced": null, "retirement": "2026-10-15", "replacement": null, "checked": "2026-11-03", "sourceType": "catalog", "sourceUrl": "https://openrouter.ai/unbiased/pareto-26.10-preview", "note": "The October 16 OpenRouter catalog no longer lists this preview. The provider released stable 26.10 on October 15 under native ID pareto-26.10; pareto now selects it." }, { "model": "Pareto Latest on OpenRouter", "providerId": "~unbiased/pareto-latest", "status": "Moving alias \u00b7 points to Pareto 26.9", "announced": null, "retirement": null, "replacement": "unbiased/pareto-26.9", "checked": "2026-11-03", "sourceType": "catalog", "sourceUrl": "https://openrouter.ai/api/v1/models", "note": "This alias currently selects Pareto 26.9. It can move to another version without changing your request ID; use a named version when repeatability matters." }, { "model": "Gemini 4 Argon launch prices", "providerId": "gemini-4-argon", "status": "Introductory prices end November 30", "announced": "2026-10-13", "retirement": null, "replacement": null, "checked": "2026-11-03", "sourceType": "pricing", "sourceUrl": "https://ai.google.dev/gemini-api/docs/pricing", "effectiveLabel": "December 1, 2026", "note": "Standard input/cache/output prices per million tokens change from $2/$0.10/$10 to $4/$0.20/$20 on December 1. Batch, Flex and Priority have separate rates; thinking is billed as output." }, { "model": "Reflection Beam API name", "providerId": "Beam-1.0-501B-A23B", "status": "Documented model ID changed", "announced": null, "retirement": null, "replacement": null, "checked": "2026-11-03", "sourceUrl": "https://developers.reflection.ai/models", "note": "Current examples use Beam-1.0-501B-A23B instead of Beam-501B-A23B. Check the models available to your account. A new spelling is not evidence of a new model checkpoint or a retirement date." }, { "model": "DeepSeek Flash native aliases", "date": "2026-10-17", "checked": "2026-11-03", "status": "Version changed", "sourceUrl": "https://api-docs.deepseek.com/quick_start/pricing/", "note": "Native deepseek-flash and the retired V4 Flash names now run DeepSeek-V4.1-Flash-1017. Earlier benchmark results remain attached to the versions that were tested.", "providerId": "deepseek-flash", "announced": "2026-10-17", "retirement": null, "replacement": "DeepSeek-V4.1-Flash-1017" }, { "model": "Gemini 2.5 Pro", "providerId": "gemini-2.5-pro", "status": "Deprecated; automatic redirect scheduled", "announced": "2026-11-02", "retirement": "2026-12-01", "replacement": "gemini-3.8-flash", "checked": "2026-11-03", "sourceUrl": "https://ai.google.dev/gemini-api/docs/changelog", "note": "Google will automatically serve the replacement and charge its rates from December 1. Existing-user access remains until then; AI Studio selection has ended. Test tool calls, thinking settings and sampling compatibility before migration. Historical scores remain attached to the old version." }, { "model": "Gemini 2.5 Flash", "providerId": "gemini-2.5-flash", "status": "Deprecated; automatic redirect scheduled", "announced": "2026-11-02", "retirement": "2026-12-01", "replacement": "gemini-3.5-flash-lite", "checked": "2026-11-03", "sourceUrl": "https://ai.google.dev/gemini-api/docs/changelog", "note": "Google will automatically serve the replacement and charge its rates from December 1. Existing-user access remains until then; AI Studio selection has ended. Test tool calls, thinking settings and sampling compatibility before migration. Historical scores remain attached to the old version." } ], "corrections": [ { "checked": "2026-10-04", "note": "OpenRouter corrected the previous free-service FAQ and clarified that context is the serving model\u2019s limit. The routing page preserves unknown fixed prices and exact system identity.", "sourceUrl": "https://openrouter.ai/nvidia/switchyard", "modelId": "nvidia/switchyard" }, { "modelId": "mistral-large-4-high", "checked": "2026-10-08", "sourceUrl": "https://livebench.ai/table_2026_06_25.csv", "note": "LiveBench revised six task values in the existing Large 4 high row. Today\u2019s category scores use the corrected row; the October 7 download preserves its original values. This is a benchmark correction, not a new model release or effort sweep." }, { "modelId": "@cf/cloudflare/clef-flash and @cf/cloudflare/clef", "checked": "2026-10-11", "sourceUrl": "https://developers.cloudflare.com/workers-ai/models/clef-flash/", "note": "Current exact Workers AI pages correct stale profile facts: Flash is24,576tokens/$0.038input, and both Clef endpoints explicitly charge no output tokens. This is a source-backed correction to our data; a provider price-change date is unverified. Immutable prior downloads retain the old observation." } ], "serviceModes": { "checked": "2026-11-03", "announced": "2026-10-08", "sourceUrls": [ "https://developers.openai.com/api/docs/changelog", "https://developers.openai.com/api/docs/pricing", "https://developers.openai.com/api/docs/guides/ultrafast-mode", "https://developers.openai.com/api/docs/models/gpt-6.1-terra" ], "models": [ { "id": "gpt-6.1-sol", "name": "GPT-6.1 Sol", "context": 1050000, "maxOutput": 128000, "modes": [ { "id": "standard", "name": "Standard", "input": 2, "output": 10, "cached": 0.1, "write": 2.5, "longInput": 4, "longOutput": 15.0, "longCached": 0.2, "longWrite": 5.0, "regions": [ "global", "us", "eu" ], "write12h": 5.0, "longWrite12h": 10.0 }, { "id": "priority", "name": "Fast", "input": 4, "output": 20, "cached": 0.2, "write": 5.0, "longInput": 8, "longOutput": 30.0, "longCached": 0.4, "longWrite": 10.0, "regions": [ "global", "us", "eu" ], "write12h": 10.0, "longWrite12h": 20.0 }, { "id": "ultrafast", "name": "Ultrafast", "input": 12, "output": 60, "cached": 0.6, "write": 15.0, "longInput": 24, "longOutput": 90.0, "longCached": 1.2, "longWrite": 30.0, "regions": [ "global", "us", "eu" ], "write12h": 30.0, "longWrite12h": 60.0 } ], "maxInput": 922000 }, { "id": "gpt-6-astra", "name": "GPT-6 Astra", "context": 1050000, "maxOutput": 128000, "modes": [ { "id": "standard", "name": "Standard", "input": 10, "output": 50, "cached": 1, "write": 12.5, "longInput": 20, "longOutput": 75.0, "longCached": 2, "longWrite": 25.0, "regions": [ "global", "us", "eu" ], "write12h": 25.0, "longWrite12h": 50.0 }, { "id": "priority", "name": "Fast", "input": 20, "output": 100, "cached": 2, "write": 25.0, "longInput": 40, "longOutput": 150.0, "longCached": 4, "longWrite": 50.0, "regions": [ "global", "us", "eu" ], "write12h": 50.0, "longWrite12h": 100.0 }, { "id": "ultrafast", "name": "Ultrafast", "input": 60, "output": 300, "cached": 6, "write": 75.0, "longInput": 120, "longOutput": 450.0, "longCached": 12, "longWrite": 150.0, "regions": [ "global", "us" ], "write12h": 150.0, "longWrite12h": 300.0 } ], "maxInput": 922000 }, { "id": "gpt-6.1-terra", "name": "GPT-6.1 Terra", "context": 1050000, "maxOutput": 128000, "modes": [ { "id": "standard", "name": "Standard", "input": 1.0, "output": 5.0, "cached": 0.05, "write": 1.25, "longInput": 2.0, "longOutput": 7.5, "longCached": 0.1, "longWrite": 2.5, "regions": [ "global", "us", "eu" ], "write12h": 2.5, "longWrite12h": 5.0 }, { "id": "priority", "name": "Fast", "input": 2.0, "output": 10.0, "cached": 0.1, "write": 2.5, "longInput": 4.0, "longOutput": 15.0, "longCached": 0.2, "longWrite": 5.0, "regions": [ "global", "us", "eu" ], "write12h": 5.0, "longWrite12h": 10.0 } ], "maxInput": 922000 } ], "cacheSourceUrl": "https://developers.openai.com/api/docs/guides/prompt-caching" }, "endpointChanges": [ { "id": "qwen/qwen-plus-2025-07-28", "name": "Qwen: Qwen Plus 0728", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen-plus-2025-07-28", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-235b-a22b", "name": "Qwen: Qwen3 235B A22B", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-235b-a22b", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-30b-a3b-thinking-2507", "name": "Qwen: Qwen3 30B A3B Thinking 2507", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-30b-a3b-thinking-2507", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-8b", "name": "Qwen: Qwen3 8B", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-8b", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-coder-plus", "name": "Qwen: Qwen3 Coder Plus", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-coder-plus", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-max", "name": "Qwen: Qwen3 Max", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-max", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-max-thinking", "name": "Qwen: Qwen3 Max Thinking", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-max-thinking", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-vl-235b-a22b-thinking", "name": "Qwen: Qwen3 VL 235B A22B Thinking", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-vl-235b-a22b-thinking", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-vl-32b-instruct", "name": "Qwen: Qwen3 VL 32B Instruct", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-vl-32b-instruct", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3-vl-8b-thinking", "name": "Qwen: Qwen3 VL 8B Thinking", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-vl-8b-thinking", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "qwen/qwen3.6-max-preview", "name": "Qwen: Qwen3.6 Max Preview", "provider": "Alibaba / Qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-09", "firstAbsent": "2026-10-10", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3.6-max-preview", "nativeRetirement": null, "replacement": null, "note": "This route was listed October 9 and is absent from the October 10 catalog. We have not verified a native retirement, replacement or whether an existing account can still call it. Confirm your exact provider route before changing configuration." }, { "id": "alibaba/qwen-3.6-max-preview", "name": "Qwen 3.6 Max Preview", "provider": "Alibaba / Qwen", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-11", "firstAbsent": "2026-10-12", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/qwen-3.6-max-preview", "nativeRetirement": null, "replacement": null, "note": "This Gateway ID was listed October 11 and is missing on October 12. An application may use a different ID on another service. Check access with Gateway before switching; this observation does not prove that requests fail everywhere." }, { "id": "alibaba/qwen3-235b-a22b-thinking", "name": "Qwen3 VL 235B A22B Thinking", "provider": "Alibaba / Qwen", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-11", "firstAbsent": "2026-10-12", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/qwen3-235b-a22b-thinking", "nativeRetirement": null, "replacement": null, "note": "This Gateway ID was listed October 11 and is missing on October 12. An application may use a different ID on another service. Check access with Gateway before switching; this observation does not prove that requests fail everywhere." }, { "id": "alibaba/qwen3-coder-plus", "name": "Qwen3 Coder Plus", "provider": "Alibaba / Qwen", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-11", "firstAbsent": "2026-10-12", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/qwen3-coder-plus", "nativeRetirement": null, "replacement": null, "note": "This Gateway ID was listed October 11 and is missing on October 12. An application may use a different ID on another service. Check access with Gateway before switching; this observation does not prove that requests fail everywhere." }, { "id": "alibaba/qwen3-max", "name": "Qwen3 Max", "provider": "Alibaba / Qwen", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-11", "firstAbsent": "2026-10-12", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/qwen3-max", "nativeRetirement": null, "replacement": null, "note": "This Gateway ID was listed October 11 and is missing on October 12. An application may use a different ID on another service. Check access with Gateway before switching; this observation does not prove that requests fail everywhere." }, { "id": "alibaba/qwen3-max-preview", "name": "Qwen3 Max Preview", "provider": "Alibaba / Qwen", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-11", "firstAbsent": "2026-10-12", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/qwen3-max-preview", "nativeRetirement": null, "replacement": null, "note": "This Gateway ID was listed October 11 and is missing on October 12. An application may use a different ID on another service. Check access with Gateway before switching; this observation does not prove that requests fail everywhere." }, { "id": "alibaba/qwen3-vl-235b-a22b-instruct", "name": "Qwen3 VL 235B A22B Instruct", "provider": "Alibaba / Qwen", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-11", "firstAbsent": "2026-10-12", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/qwen3-vl-235b-a22b-instruct", "nativeRetirement": null, "replacement": null, "note": "This Gateway ID was listed October 11 and is missing on October 12. An application may use a different ID on another service. Check access with Gateway before switching; this observation does not prove that requests fail everywhere." }, { "id": "alibaba/qwen3-vl-thinking", "name": "Qwen3 VL 235B A22B Thinking", "provider": "Alibaba / Qwen", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-11", "firstAbsent": "2026-10-12", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/qwen3-vl-thinking", "nativeRetirement": null, "replacement": null, "note": "This Gateway ID was listed October 11 and is missing on October 12. An application may use a different ID on another service. Check access with Gateway before switching; this observation does not prove that requests fail everywhere." }, { "id": "anthropic/claude-3-haiku", "name": "Claude 3 Haiku", "provider": "Anthropic", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-12", "firstAbsent": "2026-10-13", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/claude-3-haiku", "nativeRetirement": null, "replacement": null, "note": "This exact Gateway ID was listed October12 and is missing October13. Confirm access with Gateway; another service may still offer a differently named route." }, { "id": "anthropic/claude-opus-4", "name": "Claude Opus 4", "provider": "Anthropic", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-12", "firstAbsent": "2026-10-13", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/claude-opus-4", "nativeRetirement": null, "replacement": null, "note": "This exact Gateway ID was listed October12 and is missing October13. Confirm access with Gateway; another service may still offer a differently named route." }, { "id": "apodex/apodex-1.1-mini:free", "name": "Apodex: Apodex 1.1 Mini (free)", "provider": "apodex", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-13", "firstAbsent": "2026-10-14", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/apodex/apodex-1.1-mini:free", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 13 and is missing October 14. Confirm a currently available service before using it; this does not establish retirement of every hosted copy." }, { "id": "inclusionai/ling-3.0-flash-fin", "name": "inclusionAI: Ling 3.0 Flash Fin", "provider": "inclusionai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-13", "firstAbsent": "2026-10-14", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/inclusionai/ling-3.0-flash-fin", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 13 and is missing October 14. Confirm a currently available service before using it; this does not establish retirement of every hosted copy." }, { "id": "inclusionai/ling-3.0-flash-sante", "name": "inclusionAI: Ling 3.0 Flash Sante", "provider": "inclusionai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-13", "firstAbsent": "2026-10-14", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/inclusionai/ling-3.0-flash-sante", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 13 and is missing October 14. Confirm a currently available service before using it; this does not establish retirement of every hosted copy." }, { "id": "inclusionai/ling-3.1-flash-free", "name": "Ling 3.1 Flash (Free)", "provider": "inclusionai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-13", "firstAbsent": "2026-10-14", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/ling-3.1-flash-free", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 13 and is missing October 14. Confirm a currently available service before using it; this does not establish retirement of every hosted copy." }, { "id": "unbiased/pareto", "name": "Pareto", "provider": "unbiased", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-14", "firstAbsent": "2026-10-15", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/unbiased/pareto", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 14 and is missing October 15. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "anthropic/claude-sonnet-4", "name": "Anthropic: Claude Sonnet 4", "provider": "anthropic", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-15", "firstAbsent": "2026-10-16", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/anthropic/claude-sonnet-4", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 15 and is missing October 16. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "unbiased/pareto-26.10-preview", "name": "Pareto 26.10 Preview", "provider": "unbiased", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-15", "firstAbsent": "2026-10-16", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/unbiased/pareto-26.10-preview", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 15 and is missing October 16. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "anthropic/claude-sonnet-4", "name": "Claude Sonnet 4", "provider": "anthropic", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-15", "firstAbsent": "2026-10-16", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/claude-sonnet-4", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 15 and is missing October 16. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "quiverai/arrow-1.1", "name": "Arrow 1.1", "provider": "quiverai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-16", "firstAbsent": "2026-10-17", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/arrow-1.1", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 16 and is missing October 17. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "perplexity/pplx-decider-v1.1-27b", "name": "Decider V1.1 27B", "provider": "perplexity", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-17", "firstAbsent": "2026-10-18", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/pplx-decider-v1.1-27b", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 17 and is missing October 18. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "deepseek/deepseek-r1", "name": "DeepSeek: R1", "provider": "deepseek", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-19", "firstAbsent": "2026-10-20", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/deepseek/deepseek-r1", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 19 and is missing October 20. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "microsoft/wizardlm-2-8x22b", "name": "WizardLM-2 8x22B", "provider": "microsoft", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-19", "firstAbsent": "2026-10-20", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/microsoft/wizardlm-2-8x22b", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 19 and is missing October 20. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "sao10k/l3.3-euryale-70b", "name": "Sao10K: Llama 3.3 Euryale 70B", "provider": "sao10k", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-19", "firstAbsent": "2026-10-20", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/sao10k/l3.3-euryale-70b", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 19 and is missing October 20. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "stealth/cat-jam", "name": "Cat Jam", "provider": "stealth", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-19", "firstAbsent": "2026-10-20", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/cat-jam", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 19 and is missing October 20. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "google/gemini-2.5-flash:batch", "name": "Google: Gemini 2.5 Flash (batch)", "provider": "google", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-20", "firstAbsent": "2026-10-21", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/google/gemini-2.5-flash:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 20 and is missing October 21. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "google/gemini-2.5-pro:batch", "name": "Google: Gemini 2.5 Pro (batch)", "provider": "google", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-20", "firstAbsent": "2026-10-21", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/google/gemini-2.5-pro:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 20 and is missing October 21. Confirm access with your service; this does not establish retirement of every hosted copy." }, { "id": "google/gemini-2.5-flash-lite:batch", "name": "Google: Gemini 2.5 Flash Lite (batch)", "provider": "google", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/google/gemini-2.5-flash-lite:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "minimax/minimax-m3.1-flash:free", "name": "MiniMax: MiniMax M3.1 Flash (free)", "provider": "minimax", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/minimax/minimax-m3.1-flash:free", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-3.5-turbo", "name": "GPT-3.5 Turbo", "provider": "openai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/gpt-3.5-turbo", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-4-turbo", "name": "GPT-4 Turbo", "provider": "openai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/gpt-4-turbo", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-4.1-nano-batch", "name": "GPT-4.1 nano (Batch)", "provider": "openai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/gpt-4.1-nano-batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-4.1-nano-fast", "name": "GPT-4.1 nano (Fast)", "provider": "openai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/gpt-4.1-nano-fast", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-image-1", "name": "GPT Image 1", "provider": "openai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/gpt-image-1", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/o4-mini-fast", "name": "o4-mini (Fast)", "provider": "openai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/o4-mini-fast", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "spacexai/grok-4.1-fast-non-reasoning", "name": "Grok 4.1 Fast Non-Reasoning", "provider": "spacexai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/grok-4.1-fast-non-reasoning", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "spacexai/grok-4.1-fast-reasoning", "name": "Grok 4.1 Fast Reasoning", "provider": "spacexai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/grok-4.1-fast-reasoning", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "spacexai/grok-voice-think-fast-1.0", "name": "Grok Voice Think Fast 1.0", "provider": "spacexai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-21", "firstAbsent": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/grok-voice-think-fast-1.0", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 21 and is missing October 23. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-3.5-turbo", "name": "OpenAI: GPT-3.5 Turbo", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-23", "firstAbsent": "2026-10-25", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/gpt-3.5-turbo", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 23 and is missing October 25. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-3.5-turbo-instruct", "name": "OpenAI: GPT-3.5 Turbo Instruct", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-23", "firstAbsent": "2026-10-25", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/gpt-3.5-turbo-instruct", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 23 and is missing October 25. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-3.5-turbo:batch", "name": "OpenAI: GPT-3.5 Turbo (batch)", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-23", "firstAbsent": "2026-10-25", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/gpt-3.5-turbo:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 23 and is missing October 25. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "arcee-ai/trinity-large-thinking", "name": "Arcee AI: Trinity Large Thinking", "provider": "arcee-ai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-26", "firstAbsent": "2026-10-27", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/arcee-ai/trinity-large-thinking", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 26 and is missing October 27. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "ibm-granite/granite-4.0-h-micro", "name": "IBM: Granite 4.0 Micro", "provider": "ibm-granite", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-26", "firstAbsent": "2026-10-27", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/ibm-granite/granite-4.0-h-micro", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 26 and is missing October 27. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "mistralai/codestral-2508:batch", "name": "Mistral: Codestral 2508 (batch)", "provider": "mistralai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-26", "firstAbsent": "2026-10-27", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/mistralai/codestral-2508:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 26 and is missing October 27. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "mistralai/ministral-8b-2512:batch", "name": "Mistral: Ministral 3 8B 2512 (batch)", "provider": "mistralai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-26", "firstAbsent": "2026-10-27", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/mistralai/ministral-8b-2512:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 26 and is missing October 27. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "mistralai/mistral-large-2512:batch", "name": "Mistral: Mistral Large 3 2512 (batch)", "provider": "mistralai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-26", "firstAbsent": "2026-10-27", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/mistralai/mistral-large-2512:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 26 and is missing October 27. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "mistralai/mistral-medium-3.1:batch", "name": "Mistral: Mistral Medium 3.1 (batch)", "provider": "mistralai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-26", "firstAbsent": "2026-10-27", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/mistralai/mistral-medium-3.1:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 26 and is missing October 27. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "mistralai/mistral-small-2603:batch", "name": "Mistral: Mistral Small 4 (batch)", "provider": "mistralai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-26", "firstAbsent": "2026-10-27", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/mistralai/mistral-small-2603:batch", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 26 and is missing October 27. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "nvidia/nemotron-3-super-120b-a12b:free", "name": "NVIDIA: Nemotron 3 Super (free)", "provider": "nvidia", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-27", "firstAbsent": "2026-10-28", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/nvidia/nemotron-3-super-120b-a12b:free", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/gpt-4-turbo", "name": "OpenAI: GPT-4 Turbo", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/gpt-4-turbo", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/o1", "name": "OpenAI: o1", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/o1", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/o1-pro", "name": "OpenAI: o1-pro", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/o1-pro", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/o3-mini", "name": "OpenAI: o3 Mini", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/o3-mini", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/o3-mini-high", "name": "OpenAI: o3 Mini High", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/o3-mini-high", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/o4-mini", "name": "OpenAI: o4 Mini", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/o4-mini", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "openai/o4-mini-high", "name": "OpenAI: o4 Mini High", "provider": "openai", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/openai/o4-mini-high", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "qwen/qwen3-next-80b-a3b-thinking", "name": "Qwen: Qwen3 Next 80B A3B Thinking", "provider": "qwen", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/qwen/qwen3-next-80b-a3b-thinking", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "upstage/solar-pro-3", "name": "Upstage: Solar Pro 3", "provider": "upstage", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-28", "firstAbsent": "2026-10-30", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/upstage/solar-pro-3", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 27 and is missing October 28. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "arcee-ai/trinity-large-thinking", "name": "Trinity Large Thinking", "provider": "arcee-ai", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-30", "firstAbsent": "2026-10-31", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/trinity-large-thinking", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 30 and is missing October 31. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "poolside/laguna-s-2.1-free", "name": "Laguna S 2.1 Free", "provider": "poolside", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-30", "firstAbsent": "2026-10-31", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/laguna-s-2.1-free", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 30 and is missing October 31. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "poolside/laguna-s-2.1", "name": "Poolside: Laguna S 2.1", "provider": "poolside", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-31", "firstAbsent": "2026-11-02", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/poolside/laguna-s-2.1", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 31 and is missing November 2. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "poolside/laguna-s-2.1:free", "name": "Poolside: Laguna S 2.1 (free)", "provider": "poolside", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-31", "firstAbsent": "2026-11-02", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/poolside/laguna-s-2.1:free", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 31 and is missing November 2. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "poolside/laguna-xs-2.1", "name": "Poolside: Laguna XS 2.1", "provider": "poolside", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-31", "firstAbsent": "2026-11-02", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/poolside/laguna-xs-2.1", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 31 and is missing November 2. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "poolside/laguna-xs-2.1:free", "name": "Poolside: Laguna XS 2.1 (free)", "provider": "poolside", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-10-31", "firstAbsent": "2026-11-02", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/poolside/laguna-xs-2.1:free", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 31 and is missing November 2. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "convaiinnovations/laya-free", "name": "Laya (Free)", "provider": "convaiinnovations", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-10-31", "firstAbsent": "2026-11-02", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/laya-free", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed October 31 and is missing November 2. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "nousresearch/hermes-3-llama-3.1-405b", "name": "Nous: Hermes 3 405B Instruct", "provider": "nousresearch", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-11-02", "firstAbsent": "2026-11-03", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/nousresearch/hermes-3-llama-3.1-405b", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed November 2 and is missing November 3. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "nousresearch/hermes-3-llama-3.1-70b", "name": "Nous: Hermes 3 70B Instruct", "provider": "nousresearch", "source": "OpenRouter", "state": "Absent from current catalog", "lastSeen": "2026-11-02", "firstAbsent": "2026-11-03", "checked": "2026-11-03", "sourceUrl": "https://openrouter.ai/api/v1/models", "catalogUrl": "https://openrouter.ai/nousresearch/hermes-3-llama-3.1-70b", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed November 2 and is missing November 3. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "meituan/longcat-2.5-preview", "name": "LongCat 2.5 Preview", "provider": "meituan", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-11-02", "firstAbsent": "2026-11-03", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/longcat-2.5-preview", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed November 2 and is missing November 3. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "xiaomi/mimo-v2.5", "name": "MiMo M2.5", "provider": "xiaomi", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-11-02", "firstAbsent": "2026-11-03", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/mimo-v2.5", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed November 2 and is missing November 3. Check access with this service; absence does not establish retirement of other hosted copies." }, { "id": "xiaomi/mimo-v2.5-pro", "name": "MiMo V2.5 Pro", "provider": "xiaomi", "source": "Vercel AI Gateway", "state": "Absent from current catalog", "lastSeen": "2026-11-02", "firstAbsent": "2026-11-03", "checked": "2026-11-03", "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "catalogUrl": "https://vercel.com/ai-gateway/models/mimo-v2.5-pro", "nativeRetirement": null, "replacement": null, "note": "This exact offer was listed November 2 and is missing November 3. Check access with this service; absence does not establish retirement of other hosted copies." } ], "providerRetirements": [ { "id": "mainline-models", "name": "Legacy mainline models", "provider": "Alibaba / QwenCloud", "sourceType": "native-notice", "retirement": "2026-10-10", "effectiveAt": "2026-10-10T00:00:00+08:00", "checked": "2026-11-03", "sourceUrl": "https://docs.qwencloud.com/changelog/model-deprecations/mainline-models", "scope": "The listed QwenCloud mainline endpoints and corresponding Coding/Token Plan models.", "note": "The provider says inference ends on this date, subject to the actual change time. We have not made authenticated account calls. Recommended replacements require your own task and billing validation; they do not inherit predecessor scores.", "entries": [ { "id": "qwen3.6-max-preview", "replacement": "qwen3.7-max" }, { "id": "qwen3-max-preview", "replacement": "qwen3.7-max" }, { "id": "qwen3-max", "replacement": "qwen3.7-max" }, { "id": "qwen3-vl-flash", "replacement": "qwen3.6-flash" }, { "id": "qwen3-coder-plus", "replacement": "qwen3.7-plus" } ] }, { "id": "historical-mainline-models", "name": "Historical mainline models", "provider": "Alibaba / QwenCloud", "sourceType": "native-notice", "retirement": "2026-10-10", "effectiveAt": "2026-10-10T00:00:00+08:00", "checked": "2026-11-03", "sourceUrl": "https://docs.qwencloud.com/changelog/model-deprecations/historical-mainline-models", "scope": "The Beijing and Singapore native regions named in this notice.", "note": "The provider says inference ends on this date, subject to the actual change time. We have not made authenticated account calls. Recommended replacements require your own task and billing validation; they do not inherit predecessor scores.", "entries": [ { "id": "qwen-turbo", "replacement": "Qwen3.6/Qwen3.7 family; choose and test an exact replacement" }, { "id": "qwen-vl-max", "replacement": "Qwen3.6/Qwen3.7 family; choose and test an exact replacement" }, { "id": "qwen-vl-plus", "replacement": "Qwen3.6/Qwen3.7 family; choose and test an exact replacement" }, { "id": "qwq-plus", "replacement": "Qwen3.6/Qwen3.7 family; choose and test an exact replacement" }, { "id": "qvq-max", "replacement": "Qwen3.6/Qwen3.7 family; choose and test an exact replacement" } ] }, { "id": "snapshot-models", "name": "Legacy snapshot models", "provider": "Alibaba / QwenCloud", "sourceType": "native-notice", "retirement": "2026-10-10", "effectiveAt": "2026-10-10T00:00:00+08:00", "checked": "2026-11-03", "sourceUrl": "https://docs.qwencloud.com/changelog/model-deprecations/snapshot-models", "scope": "The listed QwenCloud hosted snapshots and corresponding Coding/Token Plan models. Third-party family names here identify QwenCloud offerings, not those providers\u2019 own retirements.", "note": "The provider says inference ends on this date, subject to the actual change time. We have not made authenticated account calls. Recommended replacements require your own task and billing validation; they do not inherit predecessor scores.", "entries": [ { "id": "deepseek-r1-distill-qwen-7b", "replacement": "qwen3.7-plus" }, { "id": "qwen3-vl-30b-a3b-instruct", "replacement": "qwen3.7-plus" }, { "id": "qwen3-30b-a3b-instruct-2507", "replacement": "qwen3.7-plus" }, { "id": "qwen3-30b-a3b", "replacement": "qwen3.7-plus" }, { "id": "qwen3-vl-30b-a3b-thinking", "replacement": "qwen3.7-plus" }, { "id": "qwen3-30b-a3b-thinking-2507", "replacement": "qwen3.7-plus" }, { "id": "deepseek-r1-distill-qwen-14b", "replacement": "qwen3.7-plus" }, { "id": "qwen3-next-80b-a3b-instruct", "replacement": "qwen3.7-plus" }, { "id": "qwen3-coder-next", "replacement": "qwen3.7-plus" }, { "id": "qwen3-next-80b-a3b-thinking", "replacement": "qwen3.7-plus" }, { "id": "qwen3-coder-30b-a3b-instruct", "replacement": "qwen3.7-plus" }, { "id": "deepseek-v3.2", "replacement": "qwen3.7-plus" }, { "id": "deepseek-v3.2-exp", "replacement": "qwen3.7-plus" }, { "id": "deepseek-r1-distill-qwen-32b", "replacement": "qwen3.7-plus" }, { "id": "deepseek-v3", "replacement": "qwen3.7-plus" }, { "id": "qwen3-235b-a22b-instruct-2507", "replacement": "qwen3.7-plus" }, { "id": "qwen3-32b", "replacement": "qwen3.7-plus" }, { "id": "qwen3-vl-235b-a22b-instruct", "replacement": "qwen3.7-plus" }, { "id": "qwen3-vl-32b-instruct", "replacement": "qwen3.7-plus" }, { "id": "qwen3-235b-a22b", "replacement": "qwen3.7-plus" }, { "id": "qwen3-vl-32b-thinking", "replacement": "qwen3.7-plus" }, { "id": "qwen3-vl-235b-a22b-thinking", "replacement": "qwen3.7-plus" }, { "id": "qwen3-235b-a22b-thinking-2507", "replacement": "qwen3.7-plus" }, { "id": "MiniMax-M2.1", "replacement": "qwen3.7-plus" }, { "id": "glm-4.7", "replacement": "qwen3.7-plus" }, { "id": "glm-4.6", "replacement": "qwen3.7-plus" }, { "id": "deepseek-v3.1", "replacement": "qwen3.7-plus" }, { "id": "deepseek-r1", "replacement": "qwen3.7-plus" }, { "id": "deepseek-r1-0528", "replacement": "qwen3.7-plus" }, { "id": "Moonshot-Kimi-K2-Instruct", "replacement": "qwen3.7-plus" }, { "id": "kimi-k2-thinking", "replacement": "qwen3.7-plus" }, { "id": "qwen3-coder-plus-2025-09-23", "replacement": "qwen3.7-plus" }, { "id": "qwen3-coder-plus-2025-07-22", "replacement": "qwen3.7-plus" }, { "id": "qwen3-coder-480b-a35b-instruct", "replacement": "qwen3.7-plus" }, { "id": "qwen3-vl-8b-instruct", "replacement": "qwen3.6-flash" }, { "id": "qwen3-8b", "replacement": "qwen3.6-flash" }, { "id": "qwen3-vl-8b-thinking", "replacement": "qwen3.6-flash" }, { "id": "qwen3-vl-flash-2026-01-22", "replacement": "qwen3.6-flash" }, { "id": "qwen3-vl-flash-2025-10-15", "replacement": "qwen3.6-flash" }, { "id": "qwen3-14b", "replacement": "qwen3.6-flash" }, { "id": "qwen3-max-2026-01-23", "replacement": "qwen3.7-max" }, { "id": "qwen3-max-2025-09-23", "replacement": "qwen3.7-max" } ] }, { "id": "minimax-m25-m27-native", "name": "MiniMax M2.5 and M2.7 migration", "provider": "MiniMax", "sourceType": "native-notice", "retirement": "2026-11-18", "checked": "2026-11-03", "sourceUrl": "https://platform.minimax.io/docs/faq/models", "scope": "MiniMax direct API, M Plan and Token Plan; third-party hosting is separate.", "note": "From October 19, new users lose access. Existing pay-as-you-go users can continue until November 18. M Plan and Token Plan requests using old names now run M3.1 Flash, with unchanged plan allowances. The page also labels these rows retired; the detailed notice gives the account-specific transition dates. Re-test behavior rather than assuming old IDs still execute old models.", "entries": [ { "id": "MiniMax-M2.5", "replacement": "MiniMax-M3.1-Flash" }, { "id": "MiniMax-M2.5-highspeed", "replacement": "MiniMax-M3.1-Flash" }, { "id": "MiniMax-M2.7", "replacement": "MiniMax-M3.1-Flash" }, { "id": "MiniMax-M2.7-highspeed", "replacement": "MiniMax-M3.1-Flash" } ] }, { "id": "fireworks-firerouter-removed", "name": "Fireworks router migration", "provider": "Fireworks", "sourceType": "native-notice", "retirement": null, "checked": "2026-11-03", "sourceUrl": "https://docs.fireworks.ai/nexus/firerouter", "scope": "Fireworks routing IDs only.", "note": "Current documentation says the former routing names are no longer available. The exact removal date is not stated.", "entries": [ { "id": "firerouter", "replacement": "auto or a supported pinned model ID" }, { "id": "firerouter/opus", "replacement": "auto or a supported pinned model ID" }, { "id": "auto-instant", "replacement": "auto or a supported pinned model ID" } ] }, { "id": "drex-v1-0-redirect", "name": "Drex 1.0 now serves Max", "provider": "Nace.AI", "sourceType": "native-notice", "retirement": "2026-10-21", "checked": "2026-11-03", "sourceUrl": "https://drex.nace.ai/docs/drex/guides/drex-max", "scope": "Native Drex API requests using drex-v1.0.", "note": "The provider documents forwarding to Max at its $0.06 per million input rate and limits. Max must be available to the caller; it supports at most 16 questions. Historical usage keeps its old price.", "entries": [ { "id": "drex-v1.0", "replacement": "drex-max" } ] }, { "id": "kimi-k2-access-phaseout", "name": "Kimi K2 access changes", "provider": "Moonshot AI", "sourceType": "native-notice", "retirement": null, "checked": "2026-11-03", "sourceUrl": "https://platform.kimi.ai/docs/guide/model-list", "scope": "Native K2-series API access.", "note": "New users lose K2-series access October 31. Existing users remain supported for now; the provider plans to retire the series by the end of December 2026 without an exact final day.", "entries": [ { "id": "K2 series", "replacement": "Check K3.1 Pro on your tasks before migrating" } ] }, { "id": "openai-gpt35-oct23", "name": "OpenAI GPT-3.5 legacy retirement", "provider": "OpenAI", "sourceType": "native-notice", "retirement": "2026-10-23", "checked": "2026-11-03", "sourceUrl": "https://developers.openai.com/api/docs/deprecations", "scope": "The exact native GPT-3.5 IDs below.", "note": "OpenAI lists October 23 shutdown for these legacy endpoints. Test the recommended replacement before switching; this is separate from any hosting-service catalog removal.", "entries": [ { "id": "gpt-3.5-turbo-0125", "replacement": "gpt-5.6-terra" }, { "id": "gpt-3.5-turbo", "replacement": "gpt-5.6-terra" }, { "id": "gpt-3.5-turbo-completions", "replacement": "gpt-5.6-terra" }, { "id": "ft-gpt-3.5-turbo", "replacement": "gpt-5.6-terra" } ] }, { "id": "anthropic-opus48-fast-nov30", "name": "Claude Opus 4.8 fast mode ends November 30", "provider": "Anthropic", "sourceType": "native-notice", "retirement": "2026-11-30", "checked": "2026-11-03", "sourceUrl": "https://platform.claude.com/docs/en/build-with-claude/fast-mode", "scope": "Only speed: fast requests to claude-opus-4-8.", "note": "Fast-mode requests will return an error after removal. Move fast-mode workloads to Opus 5.5; this notice does not retire ordinary Opus 4.8 access or change historical scores.", "entries": [ { "id": "claude-opus-4-8 with speed: fast", "replacement": "claude-opus-5-5 with speed: fast" } ] }, { "id": "mistral-oct31-deprecations", "name": "Mistral Medium 3.5 and Large 3 retirement notice", "provider": "Mistral", "sourceType": "native-notice", "retirement": "2027-04-30", "checked": "2026-11-03", "sourceUrl": "https://docs.mistral.ai/models", "scope": "The two exact native IDs below; other hosting services may differ.", "note": "Deprecated October 31, 2026, with retirement scheduled April 30, 2027. Mistral recommends Large 4. Test the replacement on your own tasks; no predecessor scores transfer.", "entries": [ { "id": "mistral-medium-3-5", "replacement": "Mistral Large 4" }, { "id": "mistral-large-2512", "replacement": "Mistral Large 4" } ] }, { "id": "google-gemini25-december-redirects", "name": "Gemini 2.5 Pro and Flash: December model switch", "provider": "Google Gemini API", "sourceType": "native-notice", "changeType": "redirect", "retirement": "2026-12-01", "checked": "2026-11-03", "sourceUrl": "https://ai.google.dev/gemini-api/docs/changelog", "scope": "Native Gemini API requests to gemini-2.5-pro and gemini-2.5-flash. This notice does not establish another host\u2019s schedule.", "note": "The old names will automatically use the replacement models and their prices. Update and test before December 1: tool-call validation changes, thinking_level replaces thinking_budget, and some sampling parameters and prefilled turns are removed. Gemini 2.5 Flash-Lite is not included.", "entries": [ { "id": "gemini-2.5-pro", "replacement": "gemini-3.8-flash" }, { "id": "gemini-2.5-flash", "replacement": "gemini-3.5-flash-lite" } ] } ], "scope": "Reviewed text models, serving offers, composite systems and specialized decision and research models discovered since the September 23 snapshot. These are distinct kinds of systems, not a count of independent base models. Discovery and catalog addition dates are not verified launch dates. Upgrade cost scenarios use selected OpenRouter offers; other documented prices are shown separately.", "publicationDate": "2026-11-03", "publicationId": "2026-11-03" }