| Benchmark | Kimi K2 (Reasoning) | o1 |
|---|---|---|
| AA Agentic Index | 47.9 | 31.1 |
| AA Intelligence | 33.5 | 24 |
| AA-LCR | 70.3 | 63.3 |
| AA-Omniscience | -21.4 | -10.6 |
| aider_polyglot | 59.1 | 61.7 |
| AIME 2025 | 94.5 | 81.7 |
| Artificial Analysis Coding Index | 34.8 | 39.7 |
| critpt | 2.6 | 0.3 |
| gdpval | 24.5 | 11.5 |
| GPQA Diamond | 84.5 | 75.7 |
| HLE | 23.9 | 7.7 |
| IFBench | 68.1 | 70.3 |
| OmniScience Accuracy | 30.9 | 34.7 |
| OmniScience Non-Hallucination | 25.8 | 30.7 |
| scicode | 44.8 | 35.8 |
| simplebench | 39.6 | 41.7 |
| SWE-bench Verified | 71.3 | 48.9 |
| Terminal-Bench Hard | 31.1 | 12.9 |
| τ²-Bench Telecom (AA run) | 93 | 62.6 |
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.