| Benchmark | Claude Opus 4.5 | GPT-4o mini |
|---|---|---|
| AA Agentic Index | 59.6 | 1 |
| AA Intelligence | 41.9 | 7 |
| ARC-AGI-2 | 37.6 | 0 |
| Artificial Analysis Coding Index | 47.8 | 11.4 |
| Fortress | 13.6 | 48.1 |
| gdpval | 47.3 | 0 |
| GPQA Diamond | 87 | 42.6 |
| HLE | 30.8 | 4.2 |
| IFBench | 58 | 31 |
| LiveCodeBench | 87 | 35.5 |
| MathVista | 80.2 | 56.7 |
| MMMU | 80.7 | 59.4 |
| MMMU-Pro | 74 | 41.5 |
| scicode | 50 | 22.9 |
| simplebench | 62 | 10.7 |
| SWE-bench Verified | 81.5 | 8.7 |
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.