| Benchmark | Claude Opus 4.8 | Gemini 1.5 Pro |
|---|---|---|
| AA Intelligence | 57.3 | 10 |
| ARC-AGI-2 | 72.1 | 0.8 |
| arena_vision | 1280 | 1179 |
| Artificial Analysis Coding Index | 74.3 | 23.6 |
| Fortress | 18.2 | 53.9 |
| GPQA Diamond | 93.6 | 59.1 |
| HLE | 57.9 | 4.9 |
| MMMU-Pro | 78.9 | 55 |
| scicode | 53.5 | 29.5 |
| simplebench | 64.8 | 27.1 |
| SWE-bench Verified | 88.6 | 34.2 |
| Video-MME | 86 | 78.6 |
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.