o1-mini thinks before it speaks — literally. It uses internal chain-of-thought reasoning to work through problems step by step before producing an answer, making it noticeably stronger at logic, math, and code than its size might suggest. The trade-off is speed and cost: that extra reasoning takes time and tokens. It's a focused, no-frills reasoner rather than a generalist conversationalist.
| Benchmark | Score | Type | Recorded |
|---|---|---|---|
| SciCode | 1.5 | main_problem_pass@1 | 5d ago |