{ "updated": "2026-10-15", "models": [ { "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": "2026-10-15", "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": "2026-10-15", "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": 262144, "sourceUrl": "https://ai-gateway.vercel.sh/v1/models", "verification": "New hosted revision", "note": "Sort support tickets, review images or score records against choices you define. This is the new v1.1 Gateway offer; its quality has not been independently compared here.", "capabilityNote": "Gateway lists text and image input for classification, routing and scoring. The provider reports improvements on its own decision tests; these are vendor results, not a general-chat score or a measured thinking-effort comparison.", "contextNote": "Gateway lists 262,144 context tokens. Its maximum question count is not confirmed. The downloadable version\u2019s supplied code has an 8,192-token limit; that is a different serving setup.", "documentedRates": { "input": 0.02, "output": 0, "cached": null, "label": "AI Gateway offer \u00b7 USD" }, "pricingNote": "Gateway lists $0.02 per million input tokens and zero output-token price. This is half the older v1 native API input quote, but billing details and quality differ; compare actual usage on the same records.", "billingNote": "Estimate uses the Gateway input-token quote. Output tokens are free; additional per-request fees and caching terms are unverified.", "accessNote": "Use the exact v1.1 Gateway ID. The existing native quickstart still names v1; its question limit and billing details are not assumed to apply to this new offer.", "checked": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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.", "pricingNote": "Separate prices have not been confirmed. Public Argon prices and its Arena result are not assigned to this version.", "checked": "2026-10-15", "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": "2026-10-15", "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; unknown names beginning drex- or nacedm- can select the current default.", "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": "2026-10-15", "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.", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "Provider documented router", "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": "A coding router that chooses Claude Opus or a Fireworks open model for each user turn. Consider it when routine coding work could run on a cheaper model, then compare completed tasks and total session bills.", "capabilityNote": "This route chooses Claude Opus or a Fireworks open model for each turn. Opus currently means Claude Opus 5.5; the open mix includes GLM 5.3 and GLM 5.3 Flash. Read the response model field to see which answered. The simpler firerouter ID can choose Opus or GPT Sol depending on credentials and client; it is not the same pinned selection.", "accessNote": "Use a Fireworks key and an Anthropic credential for Opus. Without the Anthropic credential, open models serve the turns. Routing preference 1 selects the closed model; other settings can use cheaper open models. The docs exclude accounts with data residency and Microsoft Foundry. A router is not a separately benchmarked base model.", "contextNote": "The model page rounds context to 1000K. A guaranteed exact common limit, output cap and reasoning-effort mapping are unverified.", "pricingNote": "There is no single verified token rate for this route. Open turns bill at Fireworks rates; closed turns bill to the supplied provider account, with cache reads at that serving model\u2019s rate. Vendor session savings are not an independent benchmark or a guaranteed saving.", "checked": "2026-10-15", "catalogId": null, "catalogAdded": null, "benchmarkStatus": "Matched LiveBench pending", "benchmarkSlug": null, "rates": { "input": null, "output": null, "cached": null, "write": null, "overrides": [] }, "independentEvidence": [] }, { "checked": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": 0.04, "output": 0, "cached": null, "label": "Direct decision API base rate \u00b7 USD" }, "note": "Use this hosted model when your task has a fixed set of answers, such as routing a ticket or checking a record. It accepts images directly; it is distinct from the two downloadable d1 checkpoints.", "capabilityNote": "The native /decisions/v1/systemone endpoint accepts text and images under model d1. Context, maximum questions and calibration on your own labels remain unverified.", "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 published base input rate is $0.04 per million. Each question is billed as its own prompt, including its text and all images. Images count as 1.5 tokens per 32\u00d732-pixel patch. Use the sum of billed tokens across questions in the calculator.", "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": "2026-10-15", "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": "Public beta API offer", "specialized": "decision", "context": null, "maxOutput": null, "questionLimit": null, "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, "write": 0, "label": "Direct Decisions API \u00b7 USD" }, "note": "A dedicated typed-answer API using GPT-6 Luna. It returns predicates, choices and rubric scores with probabilities; chat benchmark results do not establish decision accuracy.", "outputType": "Typed predicates, choices and rubric scores with probabilities and confidence fields.", "capabilityNote": "Use POST /v1/decisions with model gpt-6-luna. Input supports text and images. Native context and question-count caps and an effort contract are not verified in the reviewed guide.", "contextNote": "Gateway lists 1,050,000 context tokens for its route. Its zero max_tokens is typed-output metadata, not proof of zero capacity or a native API limit.", "pricingNote": "The base input rate is $0.10 per million tokens, with no separate cache-read, cache-write or output-token charge. Regional processing and long-context input premiums apply. Other GPT-6 Luna APIs use their own rates.", "billingNote": "Base input charge only; regional and long-context premiums can increase it. Cache reads/writes and output tokens have no separate charge.", "accessNote": "This is a public beta API offer, rather than a new base model. The publisher\u2019s speed comparison is not a matched independent latency result.", "discovered": "2026-10-07", "checked": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "Early access beta", "context": 262144, "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 tool-based work, currently behind beta access. Its API limits and five effort settings are documented; comparable independent results and public token prices are not verified.", "accessNote": "New accounts join a waitlist and can create keys after access is enabled. Weights and developer artifacts are promised later in October; they are not available simply because the model is announced.", "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": "Public per-token prices are not verified. Check your account contract; neither beta access nor planned open weights establishes free inference.", "discovered": "2026-10-07", "contextNote": "The beta lists 262,144 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.", "checked": "2026-10-15", "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": "2026-10-15", "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": "Newly covered TypeSafe System One decision endpoint. The Gateway announcement is dated September 16; its catalog release field is September 15, so a base-model launch date is not inferred. Independent matched decision quality and calibration are pending.", "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": "The Gateway API quotes $0.042 per million billed input tokens and zero output-token price, observed October 4. Its model page rounds the input quote to $0.04; the calculator preserves the exact catalog quote. Caching and other fees remain unverified.", "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": "Use typesafe-ai/jev through the TypeSafe evaluation endpoint, not as a free-form chat model. The contract has no free-form output-token maximum; catalog max_tokens=0 is not a zero-capacity limit.", "checked": "2026-10-15", "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": "2026-10-15", "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": "Typed decision model hosted on AI Gateway. Access was announced October 1; the base checkpoint launch date and independent matched quality are unverified.", "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": "Boundless serves both standard and -free routes free through October 31, 2026. Select Boundless for this promotion. Afterward the standard route bills and the -free route stops; later prices, cache rates and exact cutoff time are unverified.", "documentedRates": { "input": 0, "output": 0, "cached": null, "label": "Boundless Gateway promotion \u00b7 USD" }, "accessNote": "Use convaiinnovations/laya or convaiinnovations/laya-free. The published zero quote is temporary and provider-specific.", "billingNote": "Temporary Boundless promotion; later rates and fees are unverified.", "checked": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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 Decisions API accepts fewer than 262,144 total input tokens, including state, images and all questions; this is an input cap, not a chat prompt-plus-output context allowance.", "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": "Multimodal decision model with Apache 2.0 weights and a dedicated Decisions API. It returns bounded probabilities rather than open-ended chat or code. Vendor-reported evaluations are separate from independent matched evidence; the provider launch date remains unverified.", "capabilityNote": "The dedicated Decisions API accepts text, JSON and images and supports 1\u2013128 named questions in a request. It returns typed probabilities; no chat tool-calling interface or adjustable reasoning-effort sweep is established.", "pricingNote": "Direct Decisions API billing is $0.04 per million response-reported input tokens, with free output and no per-request fee. State, all questions and images count toward input; caching is unverified.", "documentedRates": { "input": 0.04, "output": 0, "cached": null, "label": "Direct Decisions API \u00b7 USD" }, "accessNote": "Public documentation accepts any Perplexity API key. This review made no private-key inference request.", "checked": "2026-10-15", "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": "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": "2026-10-15", "input": 0.075, "output": 0.225, "cached": 0.015, "note": "The separate OpenRouter offer lists $0.075 input, $0.015 cached input and $0.225 output per million tokens. These prices differ from Gateway; use the service your application calls.", "checked": "2026-10-15" }, "checked": "2026-10-15", "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": "Provider-documented preview with always-on reasoning and five effort controls; omitted API effort defaults to max. Currently available through M Plan and MiniMax Code. Launch date, public per-token pricing, measured speed and exact matched independent benchmarks remain unverified.", "defaultEffort": "max", "apiDocumentationUrls": [ "https://platform.minimax.io/docs/api-reference/text-openai-api", "https://platform.minimax.io/docs/api-reference/text-anthropic-api" ], "accessNote": "Provider API documentation currently limits this preview to M Plan and MiniMax Code. Verify your account and subscription-key entitlement; a standard pay-as-you-go key must not be assumed to have access.", "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": "2026-10-15", "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": "pareto-26-9", "catalogId": "unbiased/pareto-26.9", "name": "Pareto 26.9", "provider": "Unbiased", "released": "2026-09-17", "catalogAdded": "2026-10-13", "checked": "2026-10-15", "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": "The OpenRouter moving alias ~unbiased/pareto-latest currently points to 26.9 and can change. The native API documents only pareto, and currently accepts unknown IDs without rejecting them; a successful native response does not prove that a named version was selected.", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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.", "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.", "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": "2026-10-15", "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.", "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": "2026-10-15", "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": "2026-10-15", "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": 1000000, "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": 0.05, "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": "2026-10-15", "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": "2026-10-15", "verification": "Public preview", "note": "Preview API access is available; open weights are planned for later in October. The October 9 routed offer now declares 1,048,576 context tokens, up from 524,288, and 262,144 maximum output. The two-week 50% launch discount announced October 6 is unchanged; exact cutoff time is unverified.", "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": 0.68, "output": 2.09, "cached": 0.07, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [] }, { "slug": "switchyard", "catalogId": "nvidia/switchyard", "name": "NVIDIA Switchyard", "provider": "NVIDIA", "released": null, "catalogAdded": "2026-09-21", "checked": "2026-10-15", "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": "2026-10-15", "verification": "Composite system", "note": "Provider-documented blended system, not an independent base-model checkpoint. Preview serving behavior can change. Vendor preliminary results use different question sets and harnesses; no matched score is inferred here.", "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": 0.8, "output": 3.2, "cached": 0.03, "write": null, "overrides": [] }, "minimumCacheTokens": null, "independentEvidence": [] }, { "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "June25 data and categories unchanged; v2 current bundle ba5ba12f inspected separately, repository404 and reuse/setup unresolved. No v2 numeric import.", "url": "https://livebench.ai/table_2026_06_25.csv" }, { "source": "Arena", "status": "HTTP403; retained October13 extraction and October8 publisher cutoff.", "url": "https://arena.ai/leaderboard/text" }, { "source": "Hex DataBench", "status": "Reuse restrictions retained; historical observations unchanged.", "url": "https://hex.tech/databench/" }, { "source": "OckBench", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://ockbench.github.io/static/data/top200_model_performance.csv" }, { "source": "OpenRouter", "status": "462 routes; no new IDs; unversioned unbiased/pareto absent.", "url": "https://openrouter.ai/api/v1/models" }, { "source": "OpenAI", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://developers.openai.com/api/docs/changelog" }, { "source": "Anthropic", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://platform.claude.com/docs/en/release-notes/overview" }, { "source": "Google", "status": "Contributor private-preview pricing added with product-improvement data use; standard quotes unchanged.", "url": "https://ai.google.dev/gemini-api/docs/changelog" }, { "source": "DeepSeek", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://api-docs.deepseek.com/quick_start/pricing/" }, { "source": "xAI", "status": "October14GrokVoice3.0 speech-to-speech release noted outside text comparison; no new general text model verified.", "url": "https://docs.x.ai/developers/release-notes" }, { "source": "Alibaba/Qwen", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://huggingface.co/Qwen" }, { "source": "Z.ai", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://docs.z.ai/release-notes/new-released" }, { "source": "Moonshot", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://www.moonshot.ai/" }, { "source": "MiniMax", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://agent.minimax.io/tools/m3-1-flash-preview" }, { "source": "Meta", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://dev.meta.ai/docs/models" }, { "source": "NVIDIA", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://developer.nvidia.com/topics/ai/nemotron" }, { "source": "Thinking Machines", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://tinker-docs.thinkingmachines.ai/tinker/models/" }, { "source": "BFCL", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://gorilla.cs.berkeley.edu/data_overall.csv" }, { "source": "Vectara", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://github.com/vectara/hallucination-leaderboard" }, { "source": "Fireworks", "status": "Bare firerouter chooses Opus or Sol from credentials/client; explicit opus profile and fallback/billing clarified.", "url": "https://fireworks.ai/blog/ember-1" }, { "source": "Perceptron", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://www.perceptron.inc/blog/introducing-perceptron-mk1-5" }, { "source": "InclusionAI / Vercel", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://vercel.com/changelog/ling-3-1-flash-is-now-available-on-ai-gateway" }, { "source": "Apodex", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://platform.apodex.ai/docs/models" }, { "source": "Unbiased", "status": "Unversioned OpenRouter offer absent;26.10preview stilllisted despiteOctober15expiry. No universal retirement inferred.", "url": "https://unbiased.ai/blog/pareto-26-10-preview/" }, { "source": "Perplexity", "status": "New v1.1 Gateway offer and Apache2 local checkpoint verified separately; 262144 hosted versus8192 supplied local implementation. Vendor evaluations not imported as independent quality.", "url": "https://docs.perplexity.ai/docs/decisions/quickstart" }, { "source": "Cloudflare", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://blog.cloudflare.com/clef-decision-models/" }, { "source": "Historical matched effort studies", "status": "No new licensed matched multi-effort study verified.", "url": "https://www.stet.sh/blog/gpt-55-codex-graphql-reasoning-curve" }, { "source": "Vercel AI Gateway exact model catalog", "url": "https://ai-gateway.vercel.sh/v1/models", "status": "465 routes; new pplx-decider-v1.1-27b verified with hosted quote separate from download." }, { "source": "Convai Innovations / Vercel", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://vercel.com/changelog/laya-decision-model-now-available-on-ai-gateway-free-through-october-31" }, { "source": "Artificial Analysis", "status": "Reuse restrictions retained; no numeric import.", "url": "https://artificialanalysis.ai/terms-of-use" }, { "source": "Mistral", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://docs.mistral.ai/resources/changelogs" }, { "source": "TypeSafe AI / Vercel", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained.", "url": "https://vercel.com/changelog/typesafe-ai-jev-now-available-on-ai-gateway" }, { "source": "Vercel production usage", "status": "Licensed14completeUTCdays October1\u201314 refreshed, separate text/all populations.", "url": "https://vercel.com/docs/ai-gateway/leaderboards" }, { "source": "Reflection", "url": "https://developers.reflection.ai/models", "status": "Documented ID nowBeam-1.0-501B-A23B; logprobs output cap8192 including reasoning and plan limit. Not counted as new base model." }, { "source": "Reka", "url": "https://reka.ai/news/rho-1-collapsing-the-multimodal-stack", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Aleph Alpha", "url": "https://huggingface.co/Aleph-Alpha/Kolibri-1", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Ai2", "url": "https://allenai.org/blog/astabrief", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Hex DataBench reuse permissions", "url": "https://learn.hex.tech/docs/legal/terms-and-conditions-2020-03-27", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "StepFun", "url": "https://platform.stepfun.com/docs/zh/guides/models/step-5-preview", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Callstack", "url": "https://apex.callstack.com/models", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Liquid AI", "url": "https://www.liquid.ai/news/models", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Ai2 announcements", "url": "https://allenai.org/news", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Microsoft", "url": "https://commandline.microsoft.com/microsoft-decision-1-model-foundry/", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "FINAL-Bench", "url": "https://huggingface.co/FINAL-Bench/Darwin-27B-ZTC-v2", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "LightOn", "url": "https://huggingface.co/blog/lightonai/lightonocr-3", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "TII / Falcon", "url": "https://huggingface.co/blog/tiiuae/falcon-emirati", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Hugging Face discovery", "url": "https://huggingface.co/blog", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Fireworks FireRouter", "url": "https://fireworks.ai/blog/introducing-firerouter-with-opus", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "AionLabs native models", "url": "https://www.aionlabs.ai/docs/models/", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Additional provider feeds", "url": "https://fireworks.ai/blog", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "Nace.AI", "url": "https://console.nace.ai/docs/drex/reference/models", "status": "v2 native now131072state/262144row, refuses over-limit instead of v1.5 fallback; body cap4MiB. Prices unchanged." }, { "source": "Cloudflare native decision contracts", "url": "https://developers.cloudflare.com/workers-ai/models/", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "JetBrains Mellum", "url": "https://blog.jetbrains.com/ai/category/releases/", "status": "Primary sources checked October 15. No additional covered release verified; original measurement dates retained." }, { "source": "ScaleDown", "url": "https://docs.scaledown.ai", "status": "Native Extract now documents at most30distinctvalues perentitytype; pricing unchanged." }, { "source": "FrontiersMind", "url": "https://huggingface.co/blog/FrontiersMind/lumma-fev-27b", "status": "Lumma-Fev-27B October14 release added as local decision checkpoint; provider claims kept separate." }, { "source": "DMIS Lab", "url": "https://huggingface.co/blog/dmis-lab/ultrabert", "status": "October14 UltraBERT150M/380M andEmbed150M verified as encoders/embeddings, outside general-chat and fixed-choice decision comparison; no scores mixed." } ], "sourceHash": "d1f7d47808e64afa4b41fa565d64ed33ab686fac173bbf428e26b62f3c9ed5ea", "priceChanges": [ { "slug": "glm-5-2", "name": "GLM-5.2", "catalogId": "z-ai/glm-5.2", "previousDate": "2026-10-14", "currentDate": "2026-10-15", "previous": { "input": 0.06, "output": 7.0, "cached": 0.059 }, "current": { "input": 0.04, "output": 7.0, "cached": 0.039 }, "scheduled": false, "previousScheduleRemoved": false, "note": "Prices for this routed offer changed. Check the service you plan to use; direct-provider prices may differ." }, { "slug": "gpt-5-6-sol", "name": "GPT-5.6 Sol", "catalogId": "openai/gpt-5.6-sol", "previousDate": "2026-10-14", "currentDate": "2026-10-15", "previous": { "input": 2.0, "output": 10.0, "cached": 0.2 }, "current": { "input": 4.0, "output": 20.0, "cached": 0.4 }, "scheduled": false, "previousScheduleRemoved": false, "note": "Prices for this routed offer changed. Check the service you plan to use; direct-provider prices may differ." }, { "slug": "kimi-k3", "name": "Kimi K3", "catalogId": "moonshotai/kimi-k3", "previousDate": "2026-10-14", "currentDate": "2026-10-15", "previous": { "input": 0.73, "output": 10.0, "cached": 0.55 }, "current": { "input": 0.5, "output": 15.0, "cached": 0.28 }, "scheduled": false, "previousScheduleRemoved": false, "note": "Prices for this routed offer changed. Check the service you plan to use; direct-provider prices may differ." }, { "slug": "qwen3-8-27b", "name": "Qwen3.8 27B", "catalogId": "qwen/qwen3.8-27b", "previousDate": "2026-10-14", "currentDate": "2026-10-15", "previous": { "input": 0.04, "output": 2.3, "cached": 0.0198 }, "current": { "input": 0.24, "output": 2.2, "cached": 0.05 }, "scheduled": false, "previousScheduleRemoved": false, "note": "Prices for this routed offer changed. Check the service you plan to use; direct-provider prices may differ." }, { "slug": "glm-5-3", "name": "GLM-5.3", "catalogId": "z-ai/glm-5.3", "previousDate": "2026-10-14", "currentDate": "2026-10-15", "previous": { "input": 0.079, "output": 4.8, "cached": 0.059 }, "current": { "input": 0.1, "output": 4.8, "cached": 0.099 }, "scheduled": false, "previousScheduleRemoved": false, "note": "Prices for this routed offer changed. Check the service you plan to use; direct-provider prices may differ." }, { "slug": "glm-5-3-flash", "name": "GLM-5.3 Flash", "catalogId": "z-ai/glm-5.3-flash", "previousDate": "2026-10-14", "currentDate": "2026-10-15", "previous": { "input": 0.036, "output": 0.9, "cached": 0.0108 }, "current": { "input": 0.15, "output": 0.5, "cached": 0.03 }, "scheduled": false, "previousScheduleRemoved": false, "note": "Prices for this routed offer changed. Check the service you plan to use; direct-provider prices may differ." }, { "slug": "deepseek-v4-1-flash", "name": "DeepSeek V4.1 Flash", "catalogId": "deepseek/deepseek-v4.1-flash", "previousDate": "2026-10-14", "currentDate": "2026-10-15", "previous": { "input": 0.15, "output": 0.6, "cached": 0.003 }, "current": { "input": 0.01584, "output": 0.891, "cached": 0.0054 }, "scheduled": false, "previousScheduleRemoved": false, "note": "Prices for this routed offer changed. Check the service you plan to 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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "Expiration date reached; still catalog-listed", "announced": null, "retirement": "2026-10-15", "replacement": null, "checked": "2026-10-15", "sourceType": "catalog", "sourceUrl": "https://openrouter.ai/unbiased/pareto-26.10-preview", "note": "The October 15 catalog still lists this preview with an October 15 expiration. Confirm access before relying on it; a scheduled date alone does not prove that requests already fail." }, { "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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." } ], "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": "2026-10-15", "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" ], "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" ] }, { "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" ] }, { "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" ] } ] }, { "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" ] }, { "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" ] }, { "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" ] } ] } ] }, "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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." } ], "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": "2026-10-15", "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": "2026-10-15", "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": "2026-10-15", "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" } ] } ], "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." }