Up to four models
o3-mini vs DeepSeek-R1
On the index (GPT-4 = 100), o3-mini scores 131 and DeepSeek-R1 130. 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 and DeepSeek-R1 are too close to call (1 point apart, within the uncertainty), ahead on 1 of 1 shared core tests.
Test by testthird-party results
o3-mini
DeepSeek-R1
Every result as a table
| Test | o3-mini | DeepSeek-R1 |
|---|---|---|
| Reasoning & knowledge | ||
| GPQA Diamond (Epoch AI) | 77.0% | 71.7% |
| Fiction.LiveBench | 50.0% | 69.4% |
| SimpleBench | 22.8% | 30.9% |
| Maths | ||
| AIME-style maths (OTIS mock) | 76.9% | 53.3% |
| MATH (level 5) | 96.5% | 93.0% |
| Coding | ||
| Aider Polyglot | 60.4% | 56.9% |
| WeirdML | 43.8% | 36.5% |
Task length · METR
31 sec
Tests at least two of these models have taken. Bold rows are the core tests.