| Benchmark | Gemini 1.5 Pro | Kimi K3 |
|---|---|---|
| AA Intelligence | 10 | 60 |
| ARC-AGI-2 | 0.8 | 12.4 |
| Artificial Analysis Coding Index | 23.6 | 76.2 |
| GPQA Diamond | 59.1 | 93.5 |
| HLE | 4.9 | 56 |
| MMMU-Pro | 55 | 83.4 |
| scicode | 29.5 | 58.7 |
| simplebench | 27.1 | 60.7 |
| Video-MME | 78.6 | 90 |
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.