| Benchmark | GPT-4o mini | Kimi K2.6 |
|---|---|---|
| AA Agentic Index | 1 | 58.7 |
| AA Intelligence | 7 | 45.1 |
| Artificial Analysis Coding Index | 11.4 | 61.8 |
| gdpval | 0 | 41.4 |
| GPQA Diamond | 42.6 | 91.1 |
| HLE | 4.2 | 37.5 |
| IFBench | 31 | 76 |
| LiveCodeBench | 35.5 | 89.6 |
| MMMU-Pro | 41.5 | 80.1 |
| scicode | 22.9 | 53.5 |
| SWE-bench Verified | 8.7 | 80.2 |
| TauBench V3 - Banking | 2.9 | 23.3 |
| Terminal-Bench 2.1 | 5.6 | 65.9 |
| τ³-Bench | 2.9 | 20.6 |
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.