| Benchmark | Claude 3.5 Sonnet | Kimi K3 |
|---|---|---|
| AA Intelligence | 10 | 60 |
| Artificial Analysis Coding Index | 30.2 | 76.2 |
| GPQA Diamond | 67.2 | 93.5 |
| HLE | 3.9 | 56 |
| MMMU-Pro | 54.7 | 83.4 |
| scicode | 36.6 | 58.7 |
| simplebench | 41.4 | 60.7 |
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.