ProofBench
Writing rigorous mathematical proofs, graded for correctness.
| Category | Math |
|---|---|
| Metric | Accuracy |
| Source | Epoch AI (evaluator) |
| License | CC BY 4.0 |
| Benchmark page | epoch.ai |
| Status | Scored |
| Weight | 1.0 within Math |
| Fit | Discrimination 3.77, noise 0.79 |
Results
| # | Model | Provider | Result | Date |
|---|---|---|---|---|
| 1 | Claude Fable 5.2 Max | Anthropic | 100.0% | 2026-11-11 |
| 2 | Claude Sonnet 5.5 Max | Anthropic | 100.0% | 2026-11-11 |
| 3 | Claude Opus 5.5 Max | Anthropic | 100.0% | 2026-11-11 |
| 4 | Claude Fable 5.1 Max | Anthropic | 100.0% | 2026-11-11 |
| 5 | GPT-6 Astra | OpenAI | 99.0% | 2026-11-11 |
| 6 | GPT-6.1 Sol | OpenAI | 99.0% | 2026-11-11 |
| 7 | Gemini 4 Argon | 99.0% | 2026-11-11 | |
| 8 | Claude Opus 5 Max | Anthropic | 99.0% | 2026-11-11 |
| 9 | Claude Fable 5 Max | Anthropic | 95.0% | 2026-11-11 |
| 10 | Kimi K3.1 Pro | Moonshot AI | 91.0% | 2026-11-11 |
| 11 | Claude Haiku 5.5 Max | Anthropic | 89.0% | 2026-11-11 |
| 12 | Kimi K3 | Moonshot AI | 87.0% | 2026-11-11 |
| 13 | Muse Wonder | Meta | 84.0% | 2026-11-11 |
| 14 | GPT-6 Sol | OpenAI | 83.0% | 2026-11-11 |
| 15 | GPT-5.6 Sol Max | OpenAI | 83.0% | 2026-11-11 |
| 16 | Claude Sonnet 5 Max | Anthropic | 77.0% | 2026-11-11 |
| 17 | Hy4 preview | Tencent | 75.0% | 2026-11-11 |
| 18 | GPT-5.6 Terra Extra High | OpenAI | 74.0% | 2026-11-11 |
| 19 | MiMo-V2.6-Pro | Xiaomi | 70.0% | 2026-11-11 |
| 20 | Claude Opus 4.8 Max | Anthropic | 69.0% | 2026-11-11 |
| 21 | GPT-6 Luna | OpenAI | 64.0% | 2026-11-11 |
| 22 | MiMo-V2.6-Flash | Xiaomi | 63.0% | 2026-11-11 |
| 23 | GPT-5.6 Luna Max | OpenAI | 60.0% | 2026-11-11 |
| 24 | Gemini 3.7 Flash | 58.0% | 2026-11-11 | |
| 25 | Qwen3.8 Max | Alibaba | 58.0% | 2026-11-11 |
| 26 | Muse Spark 1.3 Max | Meta | 58.0% | 2026-11-11 |
| 27 | DeepSeek V4 Flash 0731 | DeepSeek | 56.0% | 2026-11-11 |
| 28 | GPT-5.4 Extra High | OpenAI | 56.0% | 2026-11-11 |
| 29 | Muse Spark 1.3 | Meta | 55.0% | 2026-11-11 |
| 30 | DeepSeek V4.1 Flash | DeepSeek | 54.0% | 2026-11-11 |
| 31 | Claude Opus 4.7 Max | Anthropic | 54.0% | 2026-11-11 |
| 32 | Grok 4.6 | xAI | 51.0% | 2026-11-11 |
| 33 | DeepSeek V4 Pro 0813 | DeepSeek | 50.0% | 2026-11-11 |
| 34 | Claude Opus 4.6 Max | Anthropic | 50.0% | 2026-11-11 |
| 35 | GPT-5.5 Extra High | OpenAI | 50.0% | 2026-11-11 |
| 36 | GLM-5.3 Max | Zhipu AI | 49.0% | 2026-11-11 |
| 37 | Gemini 3.8 Flash | 48.0% | 2026-11-11 | |
| 38 | Claude Sonnet 4.6 Max | Anthropic | 45.0% | 2026-11-11 |
| 39 | Muse Spark 1.2 | Meta | 43.0% | 2026-11-11 |
| 40 | Step 5 Preview | StepFun | 42.0% | 2026-11-11 |
| 41 | Muse Spark 1.1 Extra High | Meta | 39.0% | 2026-11-11 |
| 42 | Gemini 3.6 Flash High | 36.0% | 2026-11-11 | |
| 43 | Claude Opus 4.5 High | Anthropic | 36.0% | 2026-11-11 |
| 44 | GLM-5.2 Max | Zhipu AI | 35.0% | 2026-11-11 |
| 45 | Grok 4.5 High | xAI | 31.0% | 2026-11-11 |
| 46 | Gemini 3.5 Flash High | 31.0% | 2026-11-11 | |
| 47 | Grok 4.7 | xAI | 26.0% | 2026-11-11 |
| 48 | Qwen3.7 Max | Alibaba | 26.0% | 2026-11-11 |
| 49 | Gemini 3.1 Pro Preview High | 26.0% | 2026-11-11 | |
| 50 | GLM-5.1 | Zhipu AI | 22.2% | 2026-11-11 |
| 51 | MiMo-V2.5-Pro | Xiaomi | 22.0% | 2026-11-11 |
| 52 | GLM-5.3-Flash Max | Zhipu AI | 21.0% | 2026-11-11 |
| 53 | GPT-5.4 mini Extra High | OpenAI | 21.0% | 2026-11-11 |
| 54 | Gemini 3 Pro Preview High | 20.0% | 2026-11-11 | |
| 55 | Claude Sonnet 4.5 | Anthropic | 19.0% | 2026-11-11 |
| 56 | MiniMax-M3 | MiniMax | 18.0% | 2026-11-11 |
| 57 | Beam 1.0 | Reflection | 18.0% | 2026-11-11 |
| 58 | GPT-5 High | OpenAI | 18.0% | 2026-11-11 |
| 59 | Muse Spark | Meta | 17.0% | 2026-11-11 |
| 60 | Kimi K2.6 | Moonshot AI | 16.0% | 2026-11-11 |
| 61 | MiMo-V2.5 | Xiaomi | 16.0% | 2026-11-11 |
| 62 | Qwen3.8 27B | Alibaba | 16.0% | 2026-11-11 |
| 63 | DeepSeek V4 Pro Max | DeepSeek | 16.0% | 2026-11-11 |
| 64 | Gemini 3 Flash Preview High | 15.0% | 2026-11-11 | |
| 65 | GPT-5.2 Extra High | OpenAI | 15.0% | 2026-11-11 |
| 66 | Grok 4.20 (Reasoning) | xAI | 14.0% | 2026-11-11 |
| 67 | Gemini 3.5 Flash Lite High | 13.0% | 2026-11-11 | |
| 68 | GPT-5 Nano High | OpenAI | 12.0% | 2026-11-11 |
| 69 | Grok 4.3 High | xAI | 11.0% | 2026-11-11 |
| 70 | Mistral Large 4 Preview | Mistral | 10.0% | 2026-11-11 |
| 71 | GPT-5 Mini High | OpenAI | 9.0% | 2026-11-11 |
| 72 | GPT-5.1 Codex Max High | OpenAI | 9.0% | 2026-11-11 |
| 73 | Mistral Medium 3.5 High | Mistral | 9.0% | 2026-11-11 |
| 74 | DeepSeek V3.2 Non-reasoning | DeepSeek | 8.0% | 2026-11-11 |
| 75 | GLM-4.7 | Zhipu AI | 6.0% | 2026-11-11 |
| 76 | Inkling Small | Thinking Machines | 6.0% | 2026-11-11 |
| 77 | GPT-5.4 nano High | OpenAI | 5.0% | 2026-11-11 |
| 78 | Grok 4.1 Fast (Reasoning) | xAI | 4.0% | 2026-11-11 |
| 79 | MiniMax-M2.5 | MiniMax | 4.0% | 2026-11-11 |
| 80 | MiniMax-M2.7 | MiniMax | 3.0% | 2026-11-11 |
| 81 | Mercury 2.5 | Inception | 3.0% | 2026-11-11 |
| 82 | Nemotron 3 Ultra | NVIDIA | 2.0% | 2026-11-11 |
| 83 | Laguna XS.2 | Poolside | 0.0% | 2026-11-11 |
| 84 | Laguna M.1 | Poolside | 0.0% | 2026-11-11 |
| 85 | Inkling | Thinking Machines | 0.0% | 2026-11-11 |