| Benchmark | Kimi K2.6 | o3 |
|---|---|---|
| AA Agentic Index | 58.7 | 36.1 |
| AA Intelligence | 45.1 | 31.1 |
| AA-LCR | 76.7 | 73.3 |
| AA-Omniscience | 6.4 | -15.3 |
| arena_vision | 1264 | 1217 |
| Artificial Analysis Coding Index | 61.8 | 38.4 |
| browsecomp | 86.3 | 49.7 |
| critpt | 8 | 1.1 |
| frontiermath_tier_4 | 14.6 | 2.1 |
| gdpval | 41.4 | 12.8 |
| GPQA Diamond | 91.1 | 83.3 |
| HLE | 37.5 | 20.3 |
| IFBench | 76 | 71.4 |
| LiveCodeBench | 89.6 | 84.7 |
| MMLU-Pro | 87.1 | 85 |
| MMMU-Pro | 80.1 | 76.4 |
| OmniScience Accuracy | 32.8 | 38.6 |
| OmniScience Non-Hallucination | 60.7 | 12.9 |
| scicode | 53.5 | 41 |
| simpleqa_verified | 38.7 | 53 |
| strongreject | 99.8 | 97 |
| SWE-bench Verified | 80.2 | 69.1 |
| Terminal-Bench Hard | 43.9 | 37.1 |
| τ²-Bench Telecom (AA run) | 95.9 | 80.7 |
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.