| Benchmark | Claude 4 Opus | Kimi K2.6 |
|---|---|---|
| AA Intelligence | 31.7 | 45.1 |
| AA-LCR | 40 | 76.7 |
| arena_vision | 1207 | 1264 |
| Artificial Analysis Coding Index | 34 | 61.8 |
| frontiermath_tier_4 | 4.2 | 14.6 |
| GPQA Diamond | 79.6 | 91.1 |
| HLE | 12.3 | 37.5 |
| IFBench | 53.7 | 76 |
| livebench | 74.6 | 72.2 |
| LiveCodeBench | 70.4 | 89.6 |
| LiveCodeBench v6 | 47.4 | 89.6 |
| MMLU-Pro | 86.6 | 87.1 |
| OJBench | 19.6 | 60.6 |
| scicode | 40.9 | 53.5 |
| SWE-bench Verified | 79.4 | 80.2 |
| Terminal-Bench Hard | 31.1 | 43.9 |
| vectara_answer_rate | 91 | 99.7 |
| vectara_avg_summary_length | 123.2 | 116.7 |
| vectara_factual_consistency | 88 | 89.2 |
| vectara_hallucination_rate ↓ | 12 | 10.8 |
| τ²-Bench Telecom (AA run) | 73.4 | 95.9 |
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.