| Benchmark | Claude 3.5 Sonnet | Claude Sonnet 4.6 |
|---|---|---|
| AA Intelligence | 10 | 48.4 |
| arena_vision | 1146 | 1278 |
| Artificial Analysis Coding Index | 30.2 | 63 |
| frontiermath_tier_4 | 0 | 8.3 |
| GPQA Diamond | 67.2 | 89.9 |
| harmbench | 98.1 | 24.2 |
| HLE | 3.9 | 49 |
| MMMU-Pro | 54.7 | 75.6 |
| scicode | 36.6 | 47 |
| SWE-bench Verified | 50.8 | 79.6 |
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.