o3-mini vs o4-mini
Respan Index by category
Each category is scored 0 to 100, and how far ahead the winner is. Leads under a point are close to even.
| Category | o3-mini | o4-mini |
|---|---|---|
| Coding | 4.2 o3-mini | 7.6 o4-mini |
| Agents | 13.7* o3-mini | 23.2* o4-mini |
| Reasoning | 22.5 o3-mini | 27.6 o4-mini |
| Knowledge & language | 15.3 o3-mini | 19.6 o4-mini |
| Human preference | 40.7 o3-mini | 46.1* o4-mini |
* Estimated: no published scores in that category. How the Respan Index works
Benchmarks both report
o4-mini leads on 12 of the 13 benchmarks both report. Higher is better. Each score links to where it was published.
| Benchmark | o3-mini | o4-mini |
|---|---|---|
| WeirdML (Håvard Tveit Ihle) | 43.8% | 52.6% |
| AIME 2025 | 86.7% | 91.7% |
| Chess Puzzles (Epoch AI) | 15% | 22% |
| FrontierMath Tier 4 v2 (Epoch AI) | 0% | 4.9% |
| FrontierMath Tiers 1-3 v2 (Epoch AI) | 18.6% | 36.1% |
| GPQA Diamond | 74.9% | 77.6% |
| MATH Level 5 (Epoch AI) | 96.5% | 97.8% |
| Mystery Game Puzzles (Epoch AI) | 7% | 5% |
| OTIS Mock AIME 2024-2025 (Epoch AI) | 76.9% | 81.7% |
| SimpleBench (SimpleBench) | 22.8% | 38.7% |
| SimpleQA Verified | 15.3% | 19.6% |
| GSO Opt@1 (GSO) | 1.3% | 3.6% |
| MMLU | 85.9% | 90% |
FAQ
Is o3-mini better than o4-mini?
o4-mini scores 23.5 on the Respan Index (#109), ahead of o3-mini at 17.7 (#144). o4-mini leads on 12 of the 13 benchmarks both report.
Which is cheaper, o3-mini or o4-mini?
o3-mini lists at $1.1 input and $4.4 output per 1M tokens; o4-mini at $1.1 and $4.4. Both cost about the same per token at list price.
Which has the larger context window?
Both accept 200K tokens of context.