| Benchmark | Claude 3.5 Sonnet | Gemini 3.1 Pro |
|---|---|---|
| AA Intelligence | 10 | 47.7 |
| arena_vision | 1146 | 1280 |
| Artificial Analysis Coding Index | 30.2 | 68.8 |
| Chinese SimpleQA (C-SimpleQA) | 56.8 | 85.9 |
| Codeforces | 717 | 3052 |
| Fortress | 13 | 29.8 |
| frontiermath_tier_4 | 0 | 16.7 |
| GPQA Diamond | 67.2 | 94.3 |
| HLE | 3.9 | 51.4 |
| LiveCodeBench | 32.8 | 91.7 |
| LiveCodeBench v6 | 37.2 | 91.7 |
| mmlu | 88.3 | 92.6 |
| MMLU-Pro | 78 | 91 |
| MMMU-Pro | 54.7 | 83 |
| scicode | 36.6 | 59 |
| simplebench | 41.4 | 79.6 |
| SWE-bench Verified | 50.8 | 80.6 |
Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing.