| Benchmark | Gemini 1.5 Pro | GPT-5.5 |
|---|---|---|
| AA Intelligence | 10 | 56.3 |
| ARC-AGI-2 | 0.8 | 85 |
| arena_vision | 1179 | 1283 |
| Artificial Analysis Coding Index | 23.6 | 74.9 |
| Fortress | 53.9 | 16.3 |
| GPQA Diamond | 59.1 | 93.6 |
| HLE | 4.9 | 52.2 |
| MMMU-Pro | 55 | 83.2 |
| scicode | 29.5 | 56.1 |
| simplebench | 27.1 | 69 |
| SWE-bench Verified | 34.2 | 80.6 |
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.