Calibration
CritPt → BioMysteryBench (human-solvable)
BioMysteryBench (human-solvable) is estimated from CritPt with a Hill curve fitted on 5 models measured on both: y = 0.8174 + (0.8956 − 0.8174)·x^6.00 / (0.12565^6.00 + x^6.00), R² = 0.90, cross-validated error 0.6 pp. It is used for 185 estimates.
| Estimated model | CritPt | BioMysteryBench (human-solvable) | Source |
|---|---|---|---|
| A.X K2 | 8.9% | 82.6% | estimated ± 0.6 pp, low confidence |
| Apodex 1.1 | 4.6% | 81.8% | estimated ± 0.6 pp, low confidence |
| Apodex 1.1 Mini | 4.6% | 81.8% | estimated ± 0.6 pp, low confidence |
| Celeris-1 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Claude 3 Haiku | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Claude 4.1 Opus Thinking | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Claude 4 Sonnet | 1.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| Claude Fable 5 | 28.6% | 89.5% | estimated ± 0.6 pp, medium confidence |
| Claude Fable 5.1 | 29.7% | 89.5% | estimated ± 0.6 pp, medium confidence |
| Claude Haiku 5.5 | 18.9% | 88.9% | estimated ± 0.6 pp, medium confidence |
| Claude Opus 4.5 | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Claude Opus 4.5 Thinking | 4.6% | 81.8% | estimated ± 0.6 pp, low confidence |
| Claude Opus 4.6 | 2.8% | 81.7% | estimated ± 0.6 pp, low confidence |
| Claude Opus 4.6 (Adaptive) | 12.6% | 85.7% | estimated ± 0.6 pp, low confidence |
| Claude Opus 4.7 | 5.1% | 81.8% | estimated ± 0.6 pp, low confidence |
| Claude Opus 4.7 (Adaptive) | 12.0% | 85.1% | estimated ± 0.6 pp, low confidence |
| Claude Opus 4.8 | 20.9% | 89.2% | estimated ± 0.6 pp, medium confidence |
| Claude Sonnet 4.6 | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Claude Sonnet 5 | 16.9% | 88.4% | estimated ± 0.6 pp, medium confidence |
| Command A+ | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| DeepSeek-R1 | 1.4% | 81.7% | estimated ± 0.6 pp, low confidence |
| DeepSeek V3 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| DeepSeek V3 0324 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| DeepSeek V3.1 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| DeepSeek V3.1 (Reasoning) | 2.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| DeepSeek V3.2 | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| DeepSeek V4.1 Flash | 14.3% | 87.1% | estimated ± 0.6 pp, medium confidence |
| DeepSeek V4 Flash 0731 | 16.6% | 88.3% | estimated ± 0.6 pp, medium confidence |
| DeepSeek V4 Pro 0813 | 18.0% | 88.8% | estimated ± 0.6 pp, medium confidence |
| Exaone 4.0 1.2B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Exaone 4.0 32B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemini 2.5 Flash | 1.4% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemini 2.5 Pro | 2.6% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemini 3.1 Pro | 17.7% | 88.7% | estimated ± 0.6 pp, medium confidence |
| Gemini 3.5 Flash | 13.1% | 86.1% | estimated ± 0.6 pp, low confidence |
| Gemini 3.5 Flash-Lite | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemini 3.6 Flash | 10.6% | 83.8% | estimated ± 0.6 pp, low confidence |
| Gemini 3 Flash | 1.4% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemini 3 Pro | 9.1% | 82.7% | estimated ± 0.6 pp, low confidence |
| Gemini 3 Pro Deep Think | 25.7% | 89.5% | estimated ± 0.6 pp, medium confidence |
| Gemini 4 Argon | 27.1% | 89.5% | estimated ± 0.6 pp, medium confidence |
| Gemma 3 27B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemma 4 12B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemma 4 26B A4B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemma 4 31B | 1.4% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemma 4 E2B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Gemma 4 E4B | 0.6% | 81.7% | estimated ± 0.6 pp, low confidence |
| GLM-4.5-Air | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| GLM-4.6 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| GLM-4.7 | 1.7% | 81.7% | estimated ± 0.6 pp, low confidence |
| GLM-5 | 2.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| GLM-5.1 | 4.6% | 81.8% | estimated ± 0.6 pp, low confidence |
| GLM-5.2 | 20.9% | 89.2% | estimated ± 0.6 pp, medium confidence |
| GLM-5.3 | 19.1% | 89.0% | estimated ± 0.6 pp, medium confidence |
| GLM-5.3-Flash | 15.4% | 87.8% | estimated ± 0.6 pp, medium confidence |
| GLM-5-Turbo | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| GLM-5V-Turbo | 0.6% | 81.7% | estimated ± 0.6 pp, low confidence |
| GPT-4.1 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| GPT-4.1 mini | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| GPT-4.1 nano | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| GPT-4o | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| GPT-5.1 | 4.9% | 81.8% | estimated ± 0.6 pp, low confidence |
| GPT-5.1-Codex | 5.7% | 81.8% | estimated ± 0.6 pp, low confidence |
| GPT-5.1-Codex-Max | 5.7% | 81.8% | estimated ± 0.6 pp, low confidence |
| GPT-5.2 | 11.6% | 84.7% | estimated ± 0.6 pp, low confidence |
| GPT-5.2-Codex | 8.7% | 82.5% | estimated ± 0.6 pp, low confidence |
| GPT-5.3 Codex | 16.9% | 88.4% | estimated ± 0.6 pp, medium confidence |
| GPT-5.4 | 23.4% | 89.4% | estimated ± 0.6 pp, medium confidence |
| GPT-5.4 mini | 10.0% | 83.3% | estimated ± 0.6 pp, low confidence |
| GPT-5.4 nano | 9.3% | 82.8% | estimated ± 0.6 pp, low confidence |
| GPT-5.4 Pro | 30.0% | 89.5% | estimated ± 0.6 pp, medium confidence |
| GPT-5.5 | 27.1% | 89.5% | estimated ± 0.6 pp, medium confidence |
| GPT-5.5 Pro | 30.6% | 89.5% | estimated ± 0.6 pp, medium confidence |
| GPT-5.6 Luna | 20.6% | 89.2% | estimated ± 0.6 pp, medium confidence |
| GPT-5.6 Sol | 32.3% | 89.5% | estimated ± 0.6 pp, low confidence |
| GPT-5.6 Terra | 30.0% | 89.5% | estimated ± 0.6 pp, medium confidence |
| GPT-5 (high) | 5.7% | 81.8% | estimated ± 0.6 pp, low confidence |
| GPT-5 (medium) | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| GPT-6.1 Sol | 31.7% | 89.5% | estimated ± 0.6 pp, medium confidence |
| GPT-6 Astra | 31.7% | 89.5% | estimated ± 0.6 pp, medium confidence |
| GPT-6 Luna | 19.4% | 89.0% | estimated ± 0.6 pp, medium confidence |
| GPT-6 Sol | 30.9% | 89.5% | estimated ± 0.6 pp, medium confidence |
| GPT-OSS 120B | 1.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| GPT-OSS 20B | 1.4% | 81.7% | estimated ± 0.6 pp, low confidence |
| Granite-4.0-350M | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Granite-4.0-H-1B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Granite-4.0-H-350M | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Granite 4.2 30B | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Granite 4.2 3B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Granite 4.2 8B | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Grok 4 | 2.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Grok 4.1 Fast | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Grok 4.1 Fast (Reasoning) | 2.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Grok 4.3 | 8.0% | 82.2% | estimated ± 0.6 pp, low confidence |
| Grok 4.5 | 15.4% | 87.8% | estimated ± 0.6 pp, medium confidence |
| Grok 4.6 | 17.1% | 88.5% | estimated ± 0.6 pp, medium confidence |
| Grok 4.7 | 17.7% | 88.7% | estimated ± 0.6 pp, medium confidence |
| Grok 4 Fast (Reasoning) | 2.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Grok Code Fast 1 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Hy3 | 4.9% | 81.8% | estimated ± 0.6 pp, low confidence |
| Hy3 Preview | 4.9% | 81.8% | estimated ± 0.6 pp, low confidence |
| Hy4 preview | 16.9% | 88.4% | estimated ± 0.6 pp, medium confidence |
| Inkling | 5.4% | 81.8% | estimated ± 0.6 pp, low confidence |
| Inkling-Small | 8.3% | 82.3% | estimated ± 0.6 pp, low confidence |
| K-Exaone | 1.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| K-EXAONE 2.0 | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Kimi K2.6 | 8.0% | 82.2% | estimated ± 0.6 pp, low confidence |
| Kimi K2 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Kimi K2.5 | 3.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| Kimi K2.5 (Reasoning) | 3.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| Kimi K2.7 Code | 10.0% | 83.3% | estimated ± 0.6 pp, low confidence |
| Kimi K3 | 23.4% | 89.4% | estimated ± 0.6 pp, medium confidence |
| LFM2.5-2.6B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| LFM2.5-8B-A1B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| LFM2.5-VL-1.6B-Extract | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Ling 2.6 Flash | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Ling 3.0 Flash | 1.7% | 81.7% | estimated ± 0.6 pp, low confidence |
| Ling 3.0 Flash FP8 | 1.7% | 81.7% | estimated ± 0.6 pp, low confidence |
| Ling 3.0 Flash VL | 2.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Ling 3.0 Tiny | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Ling 3.1 Flash | 18.0% | 88.8% | estimated ± 0.6 pp, medium confidence |
| Llama 3.1 405B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Llama 4 Maverick | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Llama 4 Scout | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Mercury 2.5 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| MiMo-V2.5-Pro | 4.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| MiMo-V2.6-Flash | 12.0% | 85.1% | estimated ± 0.6 pp, low confidence |
| MiMo-V2.6-Pro | 26.6% | 89.5% | estimated ± 0.6 pp, medium confidence |
| MiMo-V2-Flash | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| MiMo-V2-Omni | 1.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| MiMo-V2-Pro | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| MiniCPM5-2B | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| MiniMax M2.7 | 0.6% | 81.7% | estimated ± 0.6 pp, low confidence |
| MiniMax M3 | 3.7% | 81.7% | estimated ± 0.6 pp, low confidence |
| Mistral Large 2 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Mistral Large 3 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Mistral Large 4 | 10.6% | 83.8% | estimated ± 0.6 pp, low confidence |
| Mistral Medium 3 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Mistral Medium 3.5 128B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Mistral Small 4 | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Mistral Small 4 (Reasoning) | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Muse Glimmer 30B | 2.6% | 81.7% | estimated ± 0.6 pp, low confidence |
| Muse Spark | 11.3% | 84.4% | estimated ± 0.6 pp, low confidence |
| Muse Spark 1.1 | 15.1% | 87.6% | estimated ± 0.6 pp, medium confidence |
| Muse Spark 1.2 | 17.7% | 88.7% | estimated ± 0.6 pp, medium confidence |
| Muse Spark 1.3 | 24.9% | 89.4% | estimated ± 0.6 pp, medium confidence |
| Nemotron 3.5 Lightning 30B A3B NVFP4 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Nemotron 3 Nano 30B | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Nemotron 3 Nano Omni 30B A3B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Nemotron 3 Super 100B | 3.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| Nemotron 3 Ultra | 3.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| Nemotron Ultra 253B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| North Mini Code | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Nova Pro | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| o1 | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| o3 | 1.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| Phi-4 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Quasar 438B | 9.4% | 82.9% | estimated ± 0.6 pp, low confidence |
| Qwen3.5-122B-A10B | 0.6% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.5-27B | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.5-35B-A3B | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.5 397B | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.5 397B (Reasoning) | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.6-27B | 1.1% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.6-35B-A3B | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen 3.6 Max (preview) | 3.7% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.6 Plus | 2.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.7 Max | 13.4% | 86.4% | estimated ± 0.6 pp, low confidence |
| Qwen3.7 Plus | 9.1% | 82.7% | estimated ± 0.6 pp, low confidence |
| Qwen3.8-27B | 5.4% | 81.8% | estimated ± 0.6 pp, low confidence |
| Qwen3.8-Flash-Next | 11.1% | 84.3% | estimated ± 0.6 pp, low confidence |
| Qwen3.8 Max Preview | 17.7% | 88.7% | estimated ± 0.6 pp, medium confidence |
| Qwen3 Max | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3-Omni-30B-A3B-Instruct | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Qwen3-Omni-30B-A3B-Thinking | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Sarvam 105B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Sarvam 30B | 0.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Solar Pro 2 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Solar Pro 3 | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |
| Solar Pro 4 | 5.4% | 81.8% | estimated ± 0.6 pp, low confidence |
| Step 3.7 Flash | 2.3% | 81.7% | estimated ± 0.6 pp, low confidence |
| Step 5 Preview | 20.9% | 89.2% | estimated ± 0.6 pp, medium confidence |
| Trinity-Large-Preview | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Trinity-Large-Thinking | 0.9% | 81.7% | estimated ± 0.6 pp, low confidence |
| Ultravox v0.6 Llama 3.3 70B | 0.0% | 81.7% | estimated ± 0.6 pp, low confidence |