| Benchmark | Gemini 1.5 Pro | Opus 5 |
|---|---|---|
| AA Intelligence | 10 | 63.1 |
| ARC-AGI-2 | 0.8 | 90.4 |
| Artificial Analysis Coding Index | 23.6 | 78 |
| GPQA Diamond | 59.1 | 93.7 |
| HLE | 4.9 | 64.7 |
| MMMU-Pro | 55 | 84.7 |
| scicode | 29.5 | 55.7 |
| simplebench | 27.1 | 80.6 |
| SWE-bench Verified | 34.2 | 96 |
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.