Claude 3 Opus is Anthropic's most thoughtful and deliberate model — it takes its time to reason through complex problems, excels at nuanced writing, and handles ambiguous or multi-step tasks with careful attention. It tends toward thoroughness over speed, which means responses are detailed and well-considered but come at higher latency and cost compared to its siblings. It's the colleague who writes the long, careful memo when others dash off a quick reply.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| HellaSwag | 95.4 | accuracy | 5d ago |
| SciCode | 1.5 | main_problem_pass@1 | 5d ago |
| ARC-Challenge | 96.4 | accuracy | 5d ago |
| GSM8K | 95.0 | accuracy | 5d ago |