| Benchmark | Claude 4 Opus | Opus 5 |
|---|---|---|
| AA Intelligence | 31.7 | 63.1 |
| AA-LCR | 40 | 78.7 |
| arc_agi_1 | 27 | 97.5 |
| ARC-AGI-2 | 8.6 | 90.4 |
| Artificial Analysis Coding Index | 34 | 78 |
| GPQA Diamond | 79.6 | 93.7 |
| HLE | 12.3 | 64.7 |
| scicode | 40.9 | 55.7 |
| simplebench | 58.8 | 80.6 |
| SWE-bench Verified | 79.4 | 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.