| Benchmark | Claude Sonnet 4.6 | Gemini 1.5 Pro |
|---|---|---|
| AA Intelligence | 48.4 | 10 |
| ARC-AGI-2 | 60.4 | 0.8 |
| arena_vision | 1278 | 1179 |
| Artificial Analysis Coding Index | 63 | 23.6 |
| GPQA Diamond | 89.9 | 59.1 |
| harmbench | 24.2 | 79.9 |
| HLE | 49 | 4.9 |
| MMMU-Pro | 75.6 | 55 |
| scicode | 47 | 29.5 |
| SWE-bench Verified | 79.6 | 34.2 |
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.