Find a model
Most choices start from constraints rather than a ranking: a budget, a wait you can live with, a licence you need, one thing the model has to be good at. Set those and the shortlist ranks on the skill you asked for, not on the overall index.
431 ranked models · data as of 5 Oct 2026
Good at
Costs at most
Runs at least
Answers within
Context of
And
431 models match, highest index is GPT-6 Astra at 70.2 for $20.00 per million tokens.
| # | Model | Index | Price $/M | Speed | First answer | Context |
|---|---|---|---|---|---|---|
| 1 |  GPT-6 AstramaxOpenAI | 70.2±3 | $20.00 | 54 | 342 s | 1.1M |
| 2 |  Claude Opus 5.5highAnthropic | 70.1±5 | $8.00 | 74 | 37 s | 1M |
| 3 | Claude Fable 5.1xhighAnthropic | 69.9±6 | $20.00 | 66 | 112 s | 1M |
| 4 |  Gemini 4 ArgonhighGoogle | 69.7±4 | $4.00 | – | – | 1M |
| 5 | GPT-6.1 SolmaxOpenAI | 69.5±3 | $4.00 | 63 | 291 s | 1.1M |
| 6 | Claude Fable 5thinkingAnthropic | 69.2±7 | $20.00 | 66 | 97 s | 1M |
| 7 | Claude Opus 5highAnthropic | 68.4±4 | $10.00 | 54 | 20 s | 1M |
| 8 |  Muse Spark 1.3maxMeta | 67.9±4 | $2.00 | 152 | 65 s | 1.0M |
| 9 | GPT-5.6 SolmaxOpenAI | 67.5±4 | $8.00 | 80 | 113 s | 1.1M |
| 10 | Claude Sonnet 5.5xhighAnthropic | 67.0±6 | $4.00 | 105 | 32 s | 1M |
| 11 | GPT-6 SolmaxOpenAI | 66.3±3 | $4.00 | 93 | 141 s | 1.1M |
| 12 | GPT-5.5xhighOpenAI | 65.9±4 | $11.25 | 85 | 52 s | 1.1M |
| 13 |  Kimi K3maxMoonshot AIopen | 65.6±3 | $6.00 | 34 | 64 s | 1.0M |
| 14 | GPT-5.5 ProxhighOpenAI | 65.6±9 | $67.50 | 9 | 3.20 s | 1.1M |
| 15 |  Qwen3.8 MaxmaxAlibaba | 65.0±3 | $3.00 | 39 | 54 s | 1M |
| 16 |  MiMo-V2.6-ProXiaomiopen | 64.8±5 | $0.544 | 46 | 48 s | 1.0M |
| 17 | Gemini 3.8 FlashmediumGoogle | 64.7±5 | $1.50 | 83 | 2.47 s | 1.0M |
| 18 | Muse SparkMeta | 64.4±3 | – | – | – | 262k |
| 19 |  Grok 4.6mediumSpaceXAI | 64.3±5 | $3.00 | 76 | 24 s | 500k |
| 20 | GPT-5.4xhighOpenAI | 64.2±3 | $5.63 | 89 | 193 s | 1.1M |
| 21 |  GLM 5.3maxZhipu AIopen | 64.1±3 | $2.15 | 71 | 32 s | 1M |
| 22 | GPT-5.3-CodexxhighOpenAI | 64.0±5 | $4.81 | 95 | 69 s | 400k |
| 23 | Claude Opus 4.7Anthropic | 63.8±2 | $10.00 | 48 | 1.61 s | 1M |
| 24 | GPT-5.6 TerramaxOpenAI | 63.7±5 | $4.50 | 100 | 177 s | 1.1M |
| 25 | Grok 4.7highSpaceXAI | 63.6±6 | $3.00 | 78 | 33 s | 500k |
| 26 | Muse Spark 1.1Meta | 63.5±6 | $2.00 | 220 | 3.55 s | 1.0M |
| 27 | Qwen3.8 2.4T A95BAlibabaopen | 63.5±5 | $3.00 | 40 | 53 s | 262k |
| 28 | Gemini 3.7 FlashmediumGoogle | 63.4±3 | $1.50 | 283 | 5.62 s | 1.0M |
| 29 | GPT-5.4 ProOpenAI | 63.4±2 | $67.50 | 4 | 6.09 s | 1.1M |
| 30 | Claude Opus 4.6Anthropic | 63.1±5 | $10.00 | 38 | 1.54 s | 1M |
| 31 | Gemini 3.5 FlashmediumGoogle | 63.1±3 | $3.38 | 211 | 14 s | 1.0M |
| 32 |  DeepSeek V4 PromaxDeepSeekopen | 62.9±3 | $0.990 | 107 | 43 s | 1M |
| 33 | GLM 5.2maxZhipu AIopen | 62.7±4 | $2.15 | 84 | 27 s | 1M |
| 34 | Muse Spark 1.2xhighMeta | 62.7±4 | $2.00 | 242 | 22 s | 1.0M |
| 35 | Gemini 3.1 ProGoogle | 62.6±4 | $4.50 | 117 | 25 s | 1.0M |
| 36 | Qwen3.6 MaxmaxAlibaba | 62.5±3 | $2.92 | 73 | 30 s | 262k |
| 37 | Claude Opus 4.8maxAnthropic | 62.4±5 | $10.00 | 61 | 63 s | 1M |
| 38 | GLM 5.3 FlashZhipu AIopen | 62.4±5 | $0.238 | 54 | 40 s | 1M |
| 39 | DeepSeek V4.1 FlashmaxDeepSeekopen | 61.7±3 | $0.262 | 209 | 11 s | 1M |
| 40 |  Step 5 PreviewStepFun | 61.6±5 | $1.43 | 86 | 26 s | 1M |
Category scores are on the same scale as the index: 50 is the average evaluated model and 15 points is one standard deviation. A model only appears under a skill if it has been measured on benchmarks in that category.