RANKINGS
LLM Leaderboard rankings. Compare large language models by GPQA, AIME 2025, SWE-bench, HLE, MMMLU, BrowseComp, and MMMU-Pro benchmark scores.
Showing 1–20 of 128
| # | Model | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 |  Anthropicclaude-mythos-preview | 86.6% | 94.6% | — | 93.9% | 92.7% | 86.9% | 64.7% | — |
| 2 |  Metamuse-spark | — | 89.5% | — | 77.4% | — | — | 58.4% | 80.4% |
| 3 | Anthropicclaude-opus-4-8 | — | 93.6% | — | 88.6% | — | 84.3% | 57.9% | — |
| 4 |  OpenAIgpt-5.5-pro | — | — | — | — | — | 90.1% | 57.2% | — |
| 5 | Anthropicclaude-opus-4-7 | 81.5% | 94.2% | — | 87.6% | 91.5% | 79.3% | 54.7% | — |
| 6 | Anthropicclaude-opus-4-6 | 82.5% | 91.3% | 99.8% | 80.8% | 91.1% | 84.0% | 53.1% | 77.3% |
| 7 | OpenAIgpt-5.5 | — | 93.6% | — | — | — | 84.4% | 52.2% | 83.2% |
| 8 |  Googlegemini-3.1-pro-preview | 80.9% | 94.3% | — | 80.6% | 92.6% | 85.9% | 51.4% | 80.5% |
| 9 |  xAIgrok-4-heavy | — | 88.4% | 100.0% | — | — | — | 50.7% | — |
| 10 | Anthropicclaude-sonnet-4-6 | 76.3% | 89.9% | — | 79.6% | 89.3% | 74.7% | 49.0% | 75.6% |
| 11 | Googlegemini-3-pro-preview | 81.1% | 91.9% | 100.0% | 76.2% | 91.8% | — | 45.8% | 81.0% |
| 12 | Googlegemini-3-flash-preview | 80.8% | 90.4% | 99.7% | 78.0% | 91.8% | — | 43.5% | 81.2% |
| 13 | Alibaba Cloud / Qwen Teamqwen3.7-max | — | 92.4% | — | 80.4% | 90.3% | — | 41.4% | — |
| 14 | Googlegemini-3.5-flash | — | — | — | — | — | — | 40.2% | 83.6% |
| 15 | xAIgrok-4 | — | 87.5% | 91.7% | — | — | — | 40.0% | — |
| 16 | OpenAIgpt-5.4 | — | 92.8% | — | — | — | 82.7% | 39.8% | 81.2% |
| 17 | Baiduernie-5.0 | — | 85.0% | 87.0% | — | — | — | 39.0% | — |
| 18 | OpenAIgpt-5.2-pro-2025-12-11 | — | 93.2% | 100.0% | — | — | 77.9% | 36.6% | — |
| 19 | OpenAIgpt-5.2-2025-12-11 | 77.4% | 92.4% | 100.0% | 80.0% | 89.6% | 65.8% | 34.5% | 79.5% |
| 20 | Alibaba Cloud / Qwen Teamqwen3.6-plus | 73.3% | 90.4% | — | 78.8% | 89.5% | — | 28.8% | 78.8% |