AI Agent Store/Model comparisonsBuild an AI worker ↗THE MODEL SELECTION SERIES
Writing a function and fixing a repository are different jobs. Our Coding assistant profile gives 45% weight to LiveBench agentic coding, 30% to coding tasks, 15% to reasoning and 10% to instruction following.
Data reviewed · LiveBench question set: 25 June 2026 · View every source ↗ · Download the data ↓
Coding measures contained code generation and completion. Agentic coding measures performance in a tool-using workflow across JavaScript, TypeScript and Python. A model that excels at the first may need better tools and instructions or more attempts for the second.
Repository search, patch application, terminal access and the test loop change results. Use the same testing setup for every candidate, and record the model version and reasoning setting. Our tables preserve the source’s tested configuration.
Add the codebase context sent on each turn and all billed output across planning, edits and tests. Cached context can reduce costs, but cache policies differ. The calculator intentionally shows an uncached baseline so its assumptions are visible.
Run your own tests for framework knowledge, unfamiliar dependencies, migrations, security-sensitive code and multi-file changes. Benchmarks support a shortlist; your regression suite decides whether a change is acceptable.
YOUR WORKLOAD. YOUR PRIORITIES.
Tokens are the small pieces of text a model reads or produces. Include charged thinking tokens in output.
This estimate uses your token counts and standard text API rates. It excludes cache discounts, tool fees, retries and provider premiums. If your input and output exceed a model’s listed capacity, we leave its cost and Value rating blank. A large capacity does not guarantee that it will use every detail correctly.
Shaded zone: 70.0+ points and $21 or less. Numbers mark models with strong price–quality tradeoffs in or near this zone. On the dotted line, no other shown model is cheaper with an equal or higher score, or scores higher at the same price. Your chosen measure, usage and filters change this shortlist.
Practical Value · Coding assistant
70% Agent Fit + 30% affordability. Rankings use your workload. Filtering the list does not change how scores are calculated.
THE EVIDENCE, SIDE BY SIDE
Anthropic · max
DeepSeek · max
Anthropic · max
Anthropic · source ID max; display xhigh
Meta · xhigh
Anthropic · max
Meta · max · version 1.0
Moonshot AI · source default
Anthropic · max
Alibaba · source default · version mapping unconfirmed
OpenAI · source ID max; display xhigh
OpenAI · max
| Compare | Input / outputUSD / 1M tokens | ||||||||
|---|---|---|---|---|---|---|---|---|---|
01 Claude Opus 5.5 ↗Anthropic · max | 79.4 | 60.8 | — | 83.2 | — | $80 | 1.00M | $4 / $20Direct API | |
02 DeepSeek V4.1 Flash ↗DeepSeek · maxOpen weights | 78.8 | 84.1 | — | 81.1 | — | $1.94 | 1.05M | $0.0158 / $0.891OpenRouter catalog | |
03 Claude Fable 5.1 ↗Anthropic · max | 76.7 | 54.2 | 94.42 sources | 83.4 | 1,501±6 | $200 | 1.00M | $10 / $50Direct API | |
04 Claude Fable 5 ↗Anthropic · source ID max; display xhighSource labels disagree | 74.8 | 52.9 | — | 83.0 | — | $200 | 1.00M | $10 / $50Direct API | |
05 Muse Spark 1.3 ↗Meta · xhigh | 74.4 | 67.1 | — | 81.6 | — | $21 | 1.05M | $1.25 / $4.25OpenRouter catalog | |
06 Claude Opus 5 ↗Anthropic · max | 73.8 | 54.8 | 61.12 sources | 80.1 | 1,489±5 | $100 | 1.00M | $5 / $25Direct API | |
07 Muse Wonder ↗Meta · max · version 1.0 | 73.3 | 56.5 | — | 81.2 | — | $80 | 1.05M | $4 / $20OpenRouter catalog | |
08 Kimi K3 ↗Moonshot AI · source defaultOpen weights | 73.1 | 62.6 | — | 79.2 | — | $35 | 1.05M | $0.5 / $15OpenRouter catalog | |
09 Claude Sonnet 5.5 ↗Anthropic · max | 72.2 | 60.6 | — | 75.7 | — | $40 | 1.00M | $2 / $10Direct API | |
10 Qwen3.8 Max ↗Alibaba · source default · version mapping unconfirmed | 71.6 | — | — | 78.5 | — | — | Not verified | Not verified | |
11 GPT-5.6 Sol ↗OpenAI · source ID max; display xhighSource labels disagree | 71.4 | 55.2 | — | 81.1 | — | $80 | 1.05M | $4 / $20Direct API | |
12 GPT-6 Astra ↗OpenAI · max | 71.4 | 50.5 | 44.42 sources | 82.2 | 1,475±7 | $200 | 1.05M | $10 / $50Direct API | |
13 Gemini 3.7 Flash ↗Google · high | 71.1 | 67.8 | 50.02 sources | 78.8 | 1,486±5 · preliminary | $15 | 1.05M | $0.75 / $3.75OpenRouter catalog | |
14 GLM-5.3 ↗Z.ai · source defaultOpen weights | 70.9 | 71.9 | — | 76.1 | — | $10.6 | 1.05M | $0.1 / $4.8OpenRouter catalog | |
15 Claude Sonnet 5 ↗Anthropic · xhigh | 70.6 | 59.5 | — | 76.0 | — | $40 | 1.00M | $2 / $10Direct API | |
16 Qwen3.8 Flash Next ↗Alibaba · source defaultOpen weights | 70.3 | — | — | 76.2 | — | — | Not verified | Not verified | |
17 Muse Spark 1.2 ↗Meta · xhigh | 70.1 | 64.1 | 55.62 sources | 78.0 | 1,492±8 | $21 | 1.05M | $1.25 / $4.25OpenRouter catalog | |
18 GPT-6.1 Sol ↗OpenAI · max | 70.0 | 59.1 | 55.62 sources | 81.6 | 1,484±9 | $40 | 1.05M | $2 / $10Direct API | |
19 DeepSeek V4 Flash Vision Exp ↗DeepSeek · experimental · source defaultOpen weightsPreview | 69.7 | 75.7 | — | 76.8 | — | $3.45 | 1.05M | $0.2156 / $0.6468OpenRouter catalog | |
20 Muse Spark 1.1 ↗Meta · source ID xhigh; display highSource labels disagree | 69.6 | 63.7 | — | 75.3 | — | $21 | 1.05M | $1.25 / $4.25OpenRouter catalog | |
21 Qwen3.8 27B ↗Alibaba · source defaultOpen weights | 69.6 | 73.5 | — | 75.3 | — | $6.8 | 1.00M | $0.24 / $2.2OpenRouter catalog | |
22 GPT-5.5 ↗OpenAI · xhigh | 69.5 | 50.2 | — | 80.2 | — | $110 | 1.05M | $5 / $30Direct API | |
23 Grok 4.6 ↗xAI · source default | 69.5 | 61.5 | — | 78.0 | — | $32 | 500K | $2 / $6OpenRouter catalog | |
24 GPT-6 Sol ↗OpenAI · max | 68.5 | 58.0 | — | 79.2 | — | $40 | 1.05M | $2 / $10Direct API | |
25 GPT-5.6 Terra ↗OpenAI · source ID max; display xhighSource labels disagree | 68.3 | 56.3 | — | 77.9 | — | $44 | 1.05M | $2 / $12Direct API | |
26 Mistral Large 4 ↗Mistral · highPreview | 67.9 | 68.8 | — | 71.8 | — | $10.98 | 1.05M | $0.68 / $2.09OpenRouter catalog | |
27 Claude Opus 4.8 ↗Anthropic · max | 67.9 | 50.6 | — | 76.2 | — | $100 | 1.00M | $5 / $25Direct API | |
28 GPT-5.4 ↗OpenAI · xhigh | 67.7 | 54.1 | — | 78.0 | — | $55 | 1.05M | $2.5 / $15Direct API | |
29 Gemini 3.8 Flash ↗Google · high | 67.7 | 65.5 | 55.62 sources | 75.8 | 1,497±5 · preliminary | $15 | 1.05M | $0.75 / $3.75OpenRouter catalog | |
30 DeepSeek V4 Pro 0813 ↗DeepSeek · source defaultOpen weights | 67.5 | 70.0 | — | 77.4 | — | $10.56 | 1.05M | $0.66 / $1.98OpenRouter catalog | |
31 Grok 4.7 ↗xAI · xhigh | 67.4 | 60.1 | — | 77.4 | — | $32 | 500K | $2 / $6OpenRouter catalog | |
32 Claude Haiku 5.5 ↗Anthropic · max | 67.2 | 75.3 | — | 67.9 | — | $2 | 1.00M | $0.1 / $0.5Direct API | |
33 Claude Opus 4.7 ↗Anthropic · xhigh | 67.2 | 50.1 | — | 76.5 | — | $100 | 1.00M | $5 / $25Direct API | |
34 Grok 4.5 ↗xAI · source default | 66.2 | 59.3 | — | 75.8 | — | $32 | 500K | $2 / $6OpenRouter catalog | |
35 GLM-5.3 Flash ↗Z.ai · source defaultOpen weights | 66.2 | 73.7 | — | 71.6 | — | $2.5 | 1.05M | $0.15 / $0.5OpenRouter catalog | |
36 GPT-5.2 Codex ↗OpenAI · source default | 65.6 | 53.4 | — | 74.0 | — | $45.5 | 400K | $1.75 / $14OpenRouter catalog | |
37 GPT-5.6 Luna ↗OpenAI · source ID max; display xhighSource labels disagree | 65.5 | 72.2 | — | 73.6 | — | $4.4 | 1.05M | $0.2 / $1.2Direct API | |
38 Gemini 3.5 Flash ↗Google · high | 65.4 | 57.6 | 16.72 sources | 74.6 | 1,478±4 | $33 | 1.05M | $1.5 / $9OpenRouter catalog | |
39 GLM-5.2 ↗Z.ai · source defaultOpen weights | 65.2 | 64.8 | — | 73.2 | — | $14.4 | 1.05M | $0.04 / $7OpenRouter catalog | |
40 Claude Opus 4.6 ↗Anthropic · high · adaptive thinking | 65.1 | 48.7 | 55.62 sources | 74.5 | 1,504±3 | $100 | 1.00M | $5 / $25Direct API | |
41 GPT-6 Luna ↗OpenAI · max | 64.6 | 73.4 | — | 72.0 | — | $2 | 1.05M | $0.1 / $0.5Direct API | |
42 GPT-5.2 ↗OpenAI · high · 2025-12-11 | 64.1 | 52.4 | — | 74.6 | — | $45.5 | 400K | $1.75 / $14OpenRouter catalog | |
43 Gemini 3.1 Pro Preview ↗Google · highPreview | 63.3 | 52.9 | — | 77.0 | — | $44 | 1.05M | $2 / $12OpenRouter catalog | |
44 Gemini 3.6 Flash ↗Google · high | 63.2 | 62.4 | 11.12 sources | 73.6 | 1,482±4 | $15 | 1.05M | $0.75 / $3.75OpenRouter catalog | |
45 DeepSeek V4 Flash 0731 ↗DeepSeek · source defaultOpen weights | 63.1 | 73.6 | — | 74.2 | — | $1.44 | 1.05M | $0.016 / $0.64OpenRouter catalog | |
46 Kimi K2.6 ↗Moonshot AI · thinkingOpen weights | 63.0 | 64.8 | — | 70.5 | — | $13.32 | 262K | $0.65 / $3.41OpenRouter catalog | |
47 Inkling ↗Thinking Machines · xhighOpen weights | 62.3 | 59.6 | — | 71.9 | — | $18.1 | 524K | $1 / $4.05OpenRouter catalog | |
48 Claude Sonnet 4.6 ↗Anthropic · medium · adaptive thinking | 62.0 | 49.6 | — | 73.0 | — | $60 | 1.00M | $3 / $15Direct API | |
49 Qwen3.7 Max ↗Alibaba · source default | 61.8 | 57.2 | — | 73.1 | — | $23.6 | 1.00M | $1.48 / $4.43OpenRouter catalog | |
50 GPT-5.4 nano ↗OpenAI · xhigh | 61.2 | 68.7 | — | 69.6 | — | $4.5 | 400K | $0.2 / $1.25Direct API | |
51 Kimi K2.7 Code ↗Moonshot AI · source defaultOpen weights | 60.8 | 62.2 | — | 68.4 | — | $13.41 | 262K | $0.6712 / $3.35OpenRouter catalog | |
52 Claude Opus 4.5 ↗Anthropic · high · 64K thinking | 60.0 | 45.1 | — | 72.6 | — | $100 | 200K | $5 / $25Direct API | |
53 Qwen3.6 Plus ↗Alibaba · source default | 59.3 | 65.8 | — | 68.9 | — | $7.15 | 1.00M | $0.325 / $1.95OpenRouter catalog | |
54 Gemini 3.5 Flash-Lite ↗Google · high | 58.9 | 65.0 | — | 63.9 | — | $8 | 1.05M | $0.3 / $2.5OpenRouter catalog | |
55 DeepSeek V4 Pro ↗DeepSeek · source defaultOpen weights | 58.8 | 61.3 | — | 71.6 | — | $13.37 | 1.05M | $0.9553 / $1.91OpenRouter catalog | |
56 Nemotron 3 Ultra ↗NVIDIA · 550B A55B · source defaultOpen weights | 57.2 | 61.8 | — | 67.4 | — | $10.8 | 262K | $0.6 / $2.4OpenRouter catalog | |
57 GPT-5.4 mini ↗OpenAI · xhigh | 56.9 | 56.9 | — | 66.4 | — | $16.5 | 400K | $0.75 / $4.5Direct API | |
58 MiniMax M3 ↗MiniMax · source default | 55.7 | 64.3 | — | 67.3 | — | $5.4 | 1.05M | $0.3 / $1.2OpenRouter catalog | |
59 Qwen3.6 27B ↗Alibaba · source defaultOpen weights | 55.1 | 61.8 | — | 64.0 | — | $9.6 | 262K | $0.32 / $3.2OpenRouter catalog | |
60 DeepSeek V4 Flash ↗DeepSeek · source defaultOpen weights | 54.6 | 68.2 | — | 65.5 | — | $1.27 | 1.05M | $0.091 / $0.182OpenRouter catalog | |
61 Grok 4.3 ↗xAI · source default | 46.2 | 48.9 | — | 62.2 | — | $17.5 | 1.00M | $1.25 / $2.5OpenRouter catalog |
— means a comparable result is unavailable. Consensus combines results for only 10 matched configurations. Model IDs, reasoning settings, source links and pricing differences are available on every model page. Downloadable model files can still have restrictions on commercial use.
LOOK BEYOND ONE NUMBER
| Measure | Claude Opus 5.5max | GPT-6 Solmax | DeepSeek V4.1 Flashmax |
|---|---|---|---|
| Agent Fit | 79.4 | 68.5 | 78.8 |
| Practical Value | 60.8 | 58.0 | 84.1 |
| Evidence Consensus | Insufficient evidence | Insufficient evidence | Insufficient evidence |
| Workload cost | $80 | $40 | $1.94 |
| Context window | 1.00M | 1.05M | 1.05M |
| Tool calling | Supported | Supported | Supported |
| Reasoning | 92.2 | 88.7 | 86.7 |
| Coding | 89.3 | 81.8 | 80.0 |
| Agentic coding | 71.7 | 52.9 | 77.3 |
| Mathematics | 97.1 | 96.4 | 93.3 |
| Data analysis | 80.3 | 81.2 | 79.3 |
| Language | 86.3 | 85.3 | 81.2 |
| Instruction following | 65.7 | 68.6 | 70.0 |
| Input / output per 1M | $4 / $20 | $2 / $10 | $0.0158 / $0.891 |