| Benchmark | Kimi K2.6 | Qwen2.5 Instruct 72B |
|---|---|---|
| AA Intelligence | 45.1 | 10 |
| AA-LCR | 76.7 | 20.3 |
| AA-Omniscience | 6.4 | -52.2 |
| Artificial Analysis Coding Index | 61.8 | 11.9 |
| Chinese SimpleQA (C-SimpleQA) | 75.9 | 52.2 |
| critpt | 8 | 0 |
| GPQA Diamond | 91.1 | 49.1 |
| HLE | 37.5 | 4.2 |
| IFBench | 76 | 36.9 |
| livebench | 72.2 | 52.3 |
| LiveCodeBench | 89.6 | 55.5 |
| MMLU-Pro | 87.1 | 71.6 |
| OmniScience Accuracy | 32.8 | 17.6 |
| OmniScience Non-Hallucination | 60.7 | 15.3 |
| scicode | 53.5 | 26.7 |
| SWE-bench Verified | 80.2 | 23.8 |
| Terminal-Bench Hard | 43.9 | 4.5 |
| τ²-Bench Telecom (AA run) | 95.9 | 34.5 |
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.