Calibration
Artificial Analysis Intelligence Index → C-Eval
C-Eval is estimated from Artificial Analysis Intelligence Index with a linear curve fitted on 5 models measured on both: y = 0.2988·x + 0.8558, R² = 0.86, cross-validated error 1.0 pp. It is used for 130 estimates.
| Estimated model | Artificial Analysis Intelligence Index | C-Eval | Source |
|---|---|---|---|
| Grok 4.6 | 44.3% | 98.8% | estimated ± 1.0 pp, low confidence |
| Kimi K3 | 43.6% | 98.6% | estimated ± 1.0 pp, low confidence |
| Muse Wonder | 43.6% | 98.6% | estimated ± 1.0 pp, low confidence |
| Qwen4 Max | 42.9% | 98.4% | estimated ± 1.0 pp, low confidence |
| Claude Opus 4.8 | 41.8% | 98.1% | estimated ± 1.0 pp, low confidence |
| Claude Opus 4.7 | 40.7% | 97.7% | estimated ± 1.0 pp, low confidence |
| Muse Spark 1.2 | 39.6% | 97.4% | estimated ± 1.0 pp, low confidence |
| DeepSeek V4.1 Flash 1017 | 39.5% | 97.4% | estimated ± 1.0 pp, low confidence |
| Gemini 3.7 Flash | 39.1% | 97.3% | estimated ± 1.0 pp, low confidence |
| GPT-5.4 | 39.0% | 97.2% | estimated ± 1.0 pp, low confidence |
| Grok 4.5 | 38.8% | 97.2% | estimated ± 1.0 pp, low confidence |
| GPT-5.5 | 38.4% | 97.0% | estimated ± 1.0 pp, low confidence |
| Claude Sonnet 5 | 38.2% | 97.0% | estimated ± 1.0 pp, low confidence |
| Claude Haiku 5.5 | 36.5% | 96.5% | estimated ± 1.0 pp, low confidence |
| Muse Ember | 36.0% | 96.3% | estimated ± 1.0 pp, low confidence |
| Grok 4.7 | 35.5% | 96.2% | estimated ± 1.0 pp, low confidence |
| Gemini 3.6 Flash | 34.0% | 95.7% | estimated ± 1.0 pp, low confidence |
| Muse Spark 1.1 | 33.7% | 95.7% | estimated ± 1.0 pp, low confidence |
| GLM-5.2 | 33.7% | 95.7% | estimated ± 1.0 pp, low confidence |
| Qwen3.8 Max Preview | 33.7% | 95.6% | estimated ± 1.0 pp, low confidence |
| Claude Opus 4.6 (Adaptive) | 32.0% | 95.1% | estimated ± 1.0 pp, low confidence |
| GLM-5.3 | 31.7% | 95.0% | estimated ± 1.0 pp, low confidence |
| Muse Spark | 31.3% | 94.9% | estimated ± 1.0 pp, low confidence |
| Gemini 3.8 Flash | 30.9% | 94.8% | estimated ± 1.0 pp, low confidence |
| Inkling 2 Small | 30.6% | 94.7% | estimated ± 1.0 pp, low confidence |
| GPT-5.2 | 30.5% | 94.7% | estimated ± 1.0 pp, low confidence |
| Qwen3.7 Max | 29.5% | 94.4% | estimated ± 1.0 pp, low confidence |
| Claude Opus 4.5 Thinking | 29.1% | 94.3% | estimated ± 1.0 pp, low confidence |
| MiniMax M3.1 Flash | 29.1% | 94.3% | estimated ± 1.0 pp, low confidence |
| MiMo-V2-Pro | 28.6% | 94.1% | estimated ± 1.0 pp, low confidence |
| Step 5 Preview | 28.5% | 94.1% | estimated ± 1.0 pp, low confidence |
| GPT-5.2-Codex | 28.5% | 94.1% | estimated ± 1.0 pp, low confidence |
| Qwen 3.6 Max (preview) | 28.4% | 94.1% | estimated ± 1.0 pp, low confidence |
| Gemini 3 Pro | 28.0% | 93.9% | estimated ± 1.0 pp, low confidence |
| GPT-6 Luna | 27.9% | 93.9% | estimated ± 1.0 pp, low confidence |
| MiMo-V2.6-Flash | 26.7% | 93.6% | estimated ± 1.0 pp, medium confidence |
| GLM-5-Turbo | 26.6% | 93.5% | estimated ± 1.0 pp, medium confidence |
| Claude Opus 4.6 | 26.4% | 93.5% | estimated ± 1.0 pp, medium confidence |
| Mistral Large 4 | 26.1% | 93.4% | estimated ± 1.0 pp, medium confidence |
| MiMo-V2.5-Pro | 26.0% | 93.3% | estimated ± 1.0 pp, medium confidence |
| Muse Spark Open | 24.9% | 93.0% | estimated ± 1.0 pp, medium confidence |
| GPT-5.1 | 24.7% | 93.0% | estimated ± 1.0 pp, medium confidence |
| Claude Sonnet 4.6 | 24.7% | 93.0% | estimated ± 1.0 pp, medium confidence |
| GPT-5.4 mini | 24.1% | 92.8% | estimated ± 1.0 pp, medium confidence |
| MiMo-V2-Omni | 23.9% | 92.7% | estimated ± 1.0 pp, medium confidence |
| GPT-5.1-Codex | 23.7% | 92.7% | estimated ± 1.0 pp, medium confidence |
| GPT-5.1-Codex-Max | 23.7% | 92.7% | estimated ± 1.0 pp, medium confidence |
| GLM-5V-Turbo | 23.5% | 92.6% | estimated ± 1.0 pp, medium confidence |
| GPT-5 | 23.0% | 92.4% | estimated ± 1.0 pp, medium confidence |
| Qwen3.5-27B | 22.9% | 92.4% | estimated ± 1.0 pp, medium confidence |
| Command A+ | 22.5% | 92.3% | estimated ± 1.0 pp, medium confidence |
| Grok 4 | 22.5% | 92.3% | estimated ± 1.0 pp, medium confidence |
| o3-pro | 21.9% | 92.1% | estimated ± 1.0 pp, medium confidence |
| Claude 4 Opus Thinking | 20.6% | 91.7% | estimated ± 1.0 pp, medium confidence |
| Grok 4.1 Fast (Reasoning) | 20.4% | 91.7% | estimated ± 1.0 pp, medium confidence |
| Step 3.7 Flash | 19.5% | 91.4% | estimated ± 1.0 pp, medium confidence |
| Qwen3.5-35B-A3B | 19.3% | 91.4% | estimated ± 1.0 pp, medium confidence |
| MiniMax M3 | 18.9% | 91.2% | estimated ± 1.0 pp, medium confidence |
| Claude 4.1 Opus | 18.6% | 91.1% | estimated ± 1.0 pp, medium confidence |
| Grok 4 Fast (Reasoning) | 17.9% | 90.9% | estimated ± 1.0 pp, medium confidence |
| Gemini 3 Flash | 17.9% | 90.9% | estimated ± 1.0 pp, medium confidence |
| Beam 1.0 501B-A23B | 17.7% | 90.9% | estimated ± 1.0 pp, medium confidence |
| Claude 3.7 Sonnet Thinking | 17.7% | 90.9% | estimated ± 1.0 pp, medium confidence |
| o4-mini | 16.7% | 90.6% | estimated ± 1.0 pp, medium confidence |
| Claude 4 Opus | 16.6% | 90.5% | estimated ± 1.0 pp, medium confidence |
| Claude 4 Sonnet | 16.6% | 90.5% | estimated ± 1.0 pp, medium confidence |
| MiMo-V2-Flash | 16.1% | 90.4% | estimated ± 1.0 pp, medium confidence |
| DeepSeek V3.2 | 16.0% | 90.4% | estimated ± 1.0 pp, medium confidence |
| Qwen3 Max | 15.6% | 90.2% | estimated ± 1.0 pp, medium confidence |
| Claude 3.7 Sonnet | 15.3% | 90.1% | estimated ± 1.0 pp, medium confidence |
| o1 | 15.2% | 90.1% | estimated ± 1.0 pp, medium confidence |
| GLM-4.6 | 14.9% | 90.0% | estimated ± 1.0 pp, medium confidence |
| K-Exaone | 14.4% | 89.9% | estimated ± 1.0 pp, medium confidence |
| Solar Pro 4 | 14.2% | 89.8% | estimated ± 1.0 pp, medium confidence |
| Grok Code Fast 1 | 14.1% | 89.8% | estimated ± 1.0 pp, medium confidence |
| Ling 2.6 Flash | 14.1% | 89.8% | estimated ± 1.0 pp, medium confidence |
| DeepSeek V3.1 | 13.7% | 89.7% | estimated ± 1.0 pp, medium confidence |
| Apodex 1.1 | 13.7% | 89.7% | estimated ± 1.0 pp, medium confidence |
| Apodex 1.1 Mini | 13.7% | 89.7% | estimated ± 1.0 pp, medium confidence |
| DeepSeek-R1 | 13.1% | 89.5% | estimated ± 1.0 pp, medium confidence |
| Inkling-Small | 12.9% | 89.4% | estimated ± 1.0 pp, medium confidence |
| Kimi K2 | 12.7% | 89.4% | estimated ± 1.0 pp, medium confidence |
| GPT-4.1 | 12.7% | 89.4% | estimated ± 1.0 pp, medium confidence |
| Inkling | 12.5% | 89.3% | estimated ± 1.0 pp, medium confidence |
| o3-mini | 12.5% | 89.3% | estimated ± 1.0 pp, medium confidence |
| o1-pro | 12.4% | 89.3% | estimated ± 1.0 pp, medium confidence |
| Gemini 3.5 Flash-Lite | 11.7% | 89.1% | estimated ± 1.0 pp, medium confidence |
| o1-preview | 11.4% | 89.0% | estimated ± 1.0 pp, medium confidence |
| Grok 4.1 Fast | 11.3% | 89.0% | estimated ± 1.0 pp, medium confidence |
| GLM-4.5-Air | 11.1% | 88.9% | estimated ± 1.0 pp, medium confidence |
| GPT-4.1 mini | 10.2% | 88.6% | estimated ± 1.0 pp, medium confidence |
| Gemini 2.5 Flash | 9.9% | 88.5% | estimated ± 1.0 pp, medium confidence |
| o1-mini | 9.8% | 88.5% | estimated ± 1.0 pp, medium confidence |
| DeepSeek V3 0324 | 9.7% | 88.5% | estimated ± 1.0 pp, medium confidence |
| GPT-4.5 | 9.6% | 88.4% | estimated ± 1.0 pp, medium confidence |
| Mistral Large 3 | 9.3% | 88.4% | estimated ± 1.0 pp, low confidence |
| Magistral Medium | 9.0% | 88.3% | estimated ± 1.0 pp, low confidence |
| Mistral Medium 3 | 9.0% | 88.3% | estimated ± 1.0 pp, low confidence |
| Claude 3.5 Haiku | 8.9% | 88.3% | estimated ± 1.0 pp, low confidence |
| Gemini 2.0 Flash | 8.9% | 88.3% | estimated ± 1.0 pp, low confidence |
| Claude 3 Opus | 8.7% | 88.2% | estimated ± 1.0 pp, low confidence |
| DeepSeek V3 | 8.5% | 88.1% | estimated ± 1.0 pp, low confidence |
| GPT-4o | 8.4% | 88.1% | estimated ± 1.0 pp, low confidence |
| DeepSeek R1 Distill Qwen 32B | 8.4% | 88.1% | estimated ± 1.0 pp, low confidence |
| Magistral Small | 8.2% | 88.0% | estimated ± 1.0 pp, low confidence |
| Gemini 1.5 Pro | 7.9% | 87.9% | estimated ± 1.0 pp, low confidence |
| GPT-4.1 nano | 7.8% | 87.9% | estimated ± 1.0 pp, low confidence |
| Solar Pro 3 | 7.8% | 87.9% | estimated ± 1.0 pp, low confidence |
| Devstral Small | 7.6% | 87.9% | estimated ± 1.0 pp, low confidence |
| Mistral Large 2 | 7.6% | 87.8% | estimated ± 1.0 pp, low confidence |
| Mistral Small 3.1 | 7.1% | 87.7% | estimated ± 1.0 pp, low confidence |
| Gemini 1.5 Flash | 7.1% | 87.7% | estimated ± 1.0 pp, low confidence |
| GPT-4 Turbo | 7.0% | 87.7% | estimated ± 1.0 pp, low confidence |
| Solar Pro 2 | 7.0% | 87.7% | estimated ± 1.0 pp, low confidence |
| Nova Pro | 7.0% | 87.7% | estimated ± 1.0 pp, low confidence |
| Qwen2.5 Coder 32B Instruct | 6.7% | 87.6% | estimated ± 1.0 pp, low confidence |
| Gemini 2.5 Flash-Lite | 6.7% | 87.6% | estimated ± 1.0 pp, low confidence |
| GPT-4o mini | 6.7% | 87.6% | estimated ± 1.0 pp, low confidence |
| DeepSeek-V2.5 | 6.6% | 87.6% | estimated ± 1.0 pp, low confidence |
| Muse Glimmer 30B | 6.3% | 87.5% | estimated ± 1.0 pp, low confidence |
| Claude 3 Haiku | 5.6% | 87.2% | estimated ± 1.0 pp, low confidence |
| Nemotron 3.5 Lightning 30B A3B NVFP4 | 5.5% | 87.2% | estimated ± 1.0 pp, low confidence |
| Gemini 1.0 Pro | 5.3% | 87.2% | estimated ± 1.0 pp, low confidence |
| Granite 4.2 8B | 5.1% | 87.1% | estimated ± 1.0 pp, low confidence |
| Gemma 3 27B | 4.9% | 87.0% | estimated ± 1.0 pp, low confidence |
| Gemma 4 E4B | 4.8% | 87.0% | estimated ± 1.0 pp, low confidence |
| Granite 4.2 3B | 4.7% | 87.0% | estimated ± 1.0 pp, low confidence |
| Mistral Medium 3.5 128B | 4.7% | 87.0% | estimated ± 1.0 pp, low confidence |
| Mistral Small 4 | 4.3% | 86.9% | estimated ± 1.0 pp, low confidence |
| North Mini Code | 3.4% | 86.6% | estimated ± 1.0 pp, low confidence |
Fitted on
| Model | Artificial Analysis Intelligence Index | C-Eval |
|---|---|---|
| Claude Opus 4.5 | 23.7% | 92.2% |
| Kimi K2.5 | 23.5% | 94.0% |
| Qwen3.6-27B | 21.4% | 91.4% |
| Qwen3.6 Plus | 27.0% | 93.3% |
| QwQ-32B | 9.5% | 88.4% |