| Benchmark | Claude 3.5 Sonnet | GPT-5.6 Sol |
|---|---|---|
| AA Intelligence | 10 | 61 |
| Artificial Analysis Coding Index | 30.2 | 78.3 |
| GPQA Diamond | 67.2 | 94.6 |
| HLE | 3.9 | 49.5 |
| longbench_v2 | 41 | 67.1 |
| MMMU-Pro | 54.7 | 84.6 |
| scicode | 36.6 | 56.9 |
| simplebench | 41.4 | 64.8 |
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.