Up to four models
o3-mini vs o1-mini
On the index (GPT-4 = 100), o3-mini scores 131 and o1-mini 123. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.
o3-mini leads o1-mini by 8 points, ahead on 2 of 2 shared core tests.
Test by testthird-party results
o3-minio1-mini
Every result as a table
| Test | o3-mini | o1-mini |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 77.0% | 62.4% |
| SimpleBench | 22.8% | 18.1% |
| Maths | ||
| AIME-style maths (OTIS mock) | 76.9% | 46.9% |
| MATH (level 5) | 96.5% | 89.2% |
| FrontierMath (2025 set) | 12.4% | 1.7% |
| Coding | ||
| Aider Polyglot | 60.4% | 32.9% |
| WeirdML | 43.8% | 36.3% |
| Novel problems | ||
| ARC-AGI-2 | 3.0% | 0.8% |
| ARC-AGI-1 | 34.5% | 14.0% |
Tests at least two of these models have taken. Bold rows are the core tests.