FrontierMath Tiers 1-3
Unpublished research-level math problems with automatically checkable answers. Run by Epoch AI on its private set.
| Category | Math |
|---|---|
| Metric | Accuracy |
| Source | Epoch AI (evaluator) |
| License | CC BY 4.0 |
| Benchmark page | epoch.ai |
| Status | Scored |
| Weight | 1.3 within Math |
| Fit | Discrimination 1.98, noise 0.21 |
| Lineage | Shares weight with frontiermath |
Results
| # | Model | Provider | Result | Date |
|---|---|---|---|---|
| 1 | Claude Fable 5.2 Max | Anthropic | 95.1% | 2026-11-06 |
| 2 | GPT-6.1 Sol Max | OpenAI | 93.7% | 2026-09-29 |
| 3 | GPT-6 Astra Max | OpenAI | 93.7% | 2026-08-30 |
| 4 | Claude Opus 5.5 Max | Anthropic | 91.2% | 2026-09-22 |
| 5 | GPT-6.1 Terra Max | OpenAI | 90.9% | 2026-10-26 |
| 6 | Gemini 4 Carbon High | 90.5% | 2026-11-02 | |
| 7 | Claude Fable 5.1 Max | Anthropic | 90.2% | 2026-09-01 |
| 8 | GPT-6 Sol Max | OpenAI | 89.8% | 2026-09-22 |
| 9 | GPT-5.6 Sol Max | OpenAI | 89.1% | 2026-07-09 |
| 10 | Claude Sonnet 5.5 Max | Anthropic | 88.8% | 2026-09-29 |
| 11 | GPT-5.5 Pro Extra High | OpenAI | 87.7% | 2026-06-12 |
| 12 | Claude Fable 5 Max | Anthropic | 87.0% | 2026-06-09 |
| 13 | Gemini 4 Argon High | 86.0% | 2026-10-13 | |
| 14 | GPT-5.6 Terra Max | OpenAI | 86.0% | 2026-07-09 |
| 15 | Claude Opus 5 Max | Anthropic | 85.6% | 2026-07-24 |
| 16 | GPT-5.5 Extra High | OpenAI | 85.3% | 2026-06-11 |
| 17 | GPT-5.4 Pro Extra High | OpenAI | 82.5% | 2026-06-13 |
| 18 | GPT-5.6 Luna Max | OpenAI | 82.1% | 2026-07-09 |
| 19 | Muse Wonder Max | Meta | 82.1% | 2026-10-12 |
| 20 | Claude Opus 4.8 Max | Anthropic | 80.0% | 2026-06-10 |
| 21 | GPT-6 Luna Max | OpenAI | 79.0% | 2026-09-22 |
| 22 | GPT-5.4 Extra High | OpenAI | 78.6% | 2026-06-11 |
| 23 | Kimi K3.1 Pro Max | Moonshot AI | 78.3% | 2026-11-06 |
| 24 | Claude Haiku 5.5 Max | Anthropic | 75.1% | 2026-10-07 |
| 25 | Qwen3.8 Max Extra High | Alibaba | 74.7% | 2026-08-04 |
| 26 | Muse Spark 1.3 Extra High | Meta | 74.4% | 2026-09-16 |
| 27 | Muse Spark 1.3 Max | Meta | 74.0% | 2026-09-18 |
| 28 | GPT-5.2 Pro Extra High | OpenAI | 74.0% | 2026-06-13 |
| 29 | Kimi K3 Max | Moonshot AI | 72.2% | 2026-07-17 |
| 30 | Gemini 3.7 Flash High | 71.6% | 2026-08-14 | |
| 31 | DeepSeek V4.1 Flash 1017 Max | DeepSeek | 71.2% | 2026-10-19 |
| 32 | Claude Opus 4.7 Max | Anthropic | 70.2% | 2026-06-10 |
| 33 | Muse Spark Open Extra High | Meta | 69.8% | 2026-10-15 |
| 34 | GLM-5.3 Max | Zhipu AI | 68.8% | 2026-08-25 |
| 35 | Gemini 3.8 Flash High | 68.4% | 2026-09-02 | |
| 36 | GPT-5.2 Extra High | OpenAI | 67.4% | 2026-06-11 |
| 37 | DeepSeek V4.1 Flash Max | DeepSeek | 67.4% | 2026-10-08 |
| 38 | Grok 4.6 Extra High | xAI | 66.0% | 2026-08-14 |
| 39 | Claude Opus 4.6 Max | Anthropic | 66.0% | 2026-06-11 |
| 40 | Qwen3.8 Max 0902 Extra High | Alibaba | 65.6% | 2026-09-02 |
| 41 | Claude Sonnet 5 Max | Anthropic | 65.6% | 2026-06-30 |
| 42 | Qwen3.7 Max | Alibaba | 64.6% | 2026-06-13 |
| 43 | DeepSeek V4 Pro 0813 Max | DeepSeek | 64.6% | 2026-08-19 |
| 44 | Gemini 3.5 Flash High | 62.8% | 2026-06-10 | |
| 45 | Gemini 3.1 Pro Preview | 59.7% | 2026-06-11 | |
| 46 | GLM-5.2 Max | Zhipu AI | 59.2% | 2026-06-19 |
| 47 | Gemini 3.6 Flash High | 59.0% | 2026-08-02 | |
| 48 | Mistral Large 4 Preview High | Mistral | 58.6% | 2026-10-12 |
| 49 | DeepSeek V4 Flash 0731 Max | DeepSeek | 57.5% | 2026-08-02 |
| 50 | Kimi K2.6 | Moonshot AI | 57.2% | 2026-06-10 |
| 51 | Grok 4.5 High | xAI | 57.2% | 2026-07-09 |
| 52 | GLM-5.3-Flash Max | Zhipu AI | 55.8% | 2026-08-27 |
| 53 | GPT-5 Pro High | OpenAI | 55.8% | 2026-06-12 |
| 54 | GPT-5 High | OpenAI | 55.4% | 2026-06-10 |
| 55 | GLM-5.2 Low | Zhipu AI | 54.7% | 2026-08-29 |
| 56 | Kimi K2.7 Code | Moonshot AI | 54.0% | 2026-06-13 |
| 57 | Grok 4.7 Extra High | xAI | 53.0% | 2026-09-22 |
| 58 | Gemini 3 Flash Preview | 51.2% | 2026-06-11 | |
| 59 | GPT-5.4 mini Extra High | OpenAI | 51.2% | 2026-06-12 |
| 60 | GPT-5 Mini High | OpenAI | 46.7% | 2026-06-12 |
| 61 | Inkling Small Extra High | Thinking Machines | 46.3% | 2026-08-14 |
| 62 | DeepSeek V4 Pro Max | DeepSeek | 45.3% | 2026-06-17 |
| 63 | Grok 4.20 (Reasoning) | xAI | 44.9% | 2026-07-13 |
| 64 | GPT-5.4 nano High | OpenAI | 44.9% | 2026-06-12 |
| 65 | Grok 4.3 High | xAI | 42.8% | 2026-06-17 |
| 66 | GLM-5.2 Non-reasoning | Zhipu AI | 42.5% | 2026-08-29 |
| 67 | GPT-5.6 Luna Low | OpenAI | 41.4% | 2026-08-29 |
| 68 | GPT-5.6 Luna Non-reasoning | OpenAI | 39.7% | 2026-08-29 |
| 69 | Qwen3.6 Plus | Alibaba | 38.3% | 2026-08-30 |
| 70 | GPT-5 Low | OpenAI | 37.2% | 2026-08-27 |
| 71 | GLM-5.1 | Zhipu AI | 36.8% | 2026-08-29 |
| 72 | o4-mini High | OpenAI | 36.1% | 2026-06-11 |
| 73 | Qwen3.6 27B | Alibaba | 35.1% | 2026-08-28 |
| 74 | Qwen3.7 Plus Non-reasoning | Alibaba | 34.4% | 2026-08-29 |
| 75 | Claude Opus 4.5 Thinking | Anthropic | 34.4% | 2026-06-11 |
| 76 | Qwen3.6 27B Non-reasoning | Alibaba | 34.0% | 2026-08-29 |
| 77 | Inkling Extra High | Thinking Machines | 33.3% | 2026-08-06 |
| 78 | o3 High | OpenAI | 33.3% | 2026-08-27 |
| 79 | Qwen3.6 Plus Non-reasoning | Alibaba | 32.3% | 2026-08-29 |
| 80 | Qwen3.5 397B-A17B Non-reasoning | Alibaba | 31.2% | 2026-08-30 |
| 81 | o3 Medium | OpenAI | 29.8% | 2026-08-27 |
| 82 | Qwen3.5 397B-A17B | Alibaba | 29.5% | 2026-08-28 |
| 83 | o4-mini Medium | OpenAI | 28.8% | 2026-08-27 |
| 84 | Gemini 3.1 Flash Lite High | 27.7% | 2026-08-30 | |
| 85 | GPT-5.5 Instant | OpenAI | 26.3% | 2026-08-02 |
| 86 | Gemini 3.5 Flash Lite High | 26.0% | 2026-08-02 | |
| 87 | GLM-5.1 Non-reasoning | Zhipu AI | 24.9% | 2026-08-28 |
| 88 | Gemini 2.5 Pro | 24.6% | 2026-06-11 | |
| 89 | GPT-5.4 mini Low | OpenAI | 24.6% | 2026-08-28 |
| 90 | Claude Sonnet 4.5 Thinking | Anthropic | 23.9% | 2026-06-11 |
| 91 | Gemini 3.1 Flash Lite Low | 22.5% | 2026-08-30 | |
| 92 | Qwen3.6 Flash Non-reasoning | Alibaba | 22.5% | 2026-08-29 |
| 93 | Gemini 3.1 Flash Lite Minimal | 21.4% | 2026-08-28 | |
| 94 | Qwen3.6 35B-A3B Non-reasoning | Alibaba | 20.4% | 2026-08-30 |
| 95 | GPT-5.4 nano Low | OpenAI | 20.4% | 2026-08-28 |
| 96 | GPT-5 Nano High | OpenAI | 20.0% | 2026-06-12 |
| 97 | Qwen3.7 Flash | Alibaba | 19.3% | 2026-08-29 |
| 98 | Qwen3.7 Flash Non-reasoning | Alibaba | 19.3% | 2026-08-29 |
| 99 | o3 Low | OpenAI | 19.3% | 2026-08-27 |
| 100 | Qwen3 Max | Alibaba | 19.0% | 2026-08-30 |
| 101 | o3-mini High | OpenAI | 18.6% | 2026-06-11 |
| 102 | GPT-5 Minimal | OpenAI | 18.3% | 2026-08-27 |
| 103 | Qwen3.5 Flash Non-reasoning | Alibaba | 18.3% | 2026-08-28 |
| 104 | GPT-5 Mini Low | OpenAI | 18.3% | 2026-08-27 |
| 105 | Qwen3.6 35B-A3B | Alibaba | 17.5% | 2026-08-29 |
| 106 | Qwen3.6 Flash | Alibaba | 17.2% | 2026-08-29 |
| 107 | GPT-5.4 mini Non-reasoning | OpenAI | 17.2% | 2026-08-28 |
| 108 | o4-mini Low | OpenAI | 16.1% | 2026-08-27 |
| 109 | o1 High | OpenAI | 14.7% | 2026-08-28 |
| 110 | Claude Opus 4.1 Thinking | Anthropic | 12.6% | 2026-06-11 |
| 111 | o3-mini Medium | OpenAI | 10.5% | 2026-08-27 |
| 112 | o1 Medium | OpenAI | 10.2% | 2026-08-27 |
| 113 | Qwen3.5 Flash | Alibaba | 9.5% | 2026-08-28 |
| 114 | o1 Low | OpenAI | 8.4% | 2026-08-27 |
| 115 | GPT-4.1 mini | OpenAI | 6.7% | 2026-08-27 |
| 116 | GPT-4.1 | OpenAI | 6.0% | 2026-08-27 |
| 117 | GPT-5 Mini Minimal | OpenAI | 6.0% | 2026-08-27 |
| 118 | GPT-5 Nano Low | OpenAI | 6.0% | 2026-08-27 |
| 119 | GPT-5.4 nano Non-reasoning | OpenAI | 4.6% | 2026-08-28 |
| 120 | o3-mini Low | OpenAI | 3.9% | 2026-08-27 |
| 121 | GPT-5 Nano Minimal | OpenAI | 1.8% | 2026-08-27 |
| 122 | GPT-4o mini | OpenAI | 0.7% | 2026-08-27 |
| 123 | GPT-4 Turbo | OpenAI | 0.7% | 2026-08-27 |
| 124 | GPT-4o (2024-08-06) | OpenAI | 0.4% | 2026-08-27 |
| 125 | GPT-3.5-turbo | OpenAI | 0.0% | 2026-08-27 |