| Benchmark | Kimi K3 | o1 |
|---|---|---|
| AA Agentic Index | 54.3 | 31.1 |
| AA Intelligence | 60 | 24 |
| AA-LCR | 82.7 | 63.3 |
| AA-Omniscience | 46 | -10.6 |
| Artificial Analysis Coding Index | 76.2 | 39.7 |
| critpt | 23.4 | 0.3 |
| gdpval | 59.4 | 11.5 |
| GPQA Diamond | 93.5 | 75.7 |
| HLE | 56 | 7.7 |
| OmniScience Accuracy | 47.6 | 34.7 |
| OmniScience Non-Hallucination | 49.1 | 30.7 |
| scicode | 58.7 | 35.8 |
| simplebench | 60.7 | 41.7 |
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.