o3-mini
- Made by
- OpenAI
- Released
- Announced in
- OpenAI o3-miniopenai.com
- Event
- OpenAI o3-mini
- API record
- /api/v1/models/openai-o3-mini
Benchmark results
| Benchmark | Score | Source check |
|---|---|---|
| Humanity's Last ExamNo tools · high reasoning effort · reported later, in Introducing deep research | 13.0% | CheckedA person checked this number against its source on 20 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| high reasoning effort · reported later, in Introducing OpenAI o3 and o4-mini | 13.40% | CorroboratedThe same number is printed in 1 other lab document: 1 |
| GPQA Diamondhigh reasoning effort | 79.7% | |
| high reasoning effort; answers parsed by a model, not regex · reported later, in Introducing GPT-4.1 in the API | 77.2% | CheckedA person checked this number against its source on 20 Oct 2026. This confirms the number matches what the lab published, not that the score was reproduced. |
| high reasoning effort; no tools · reported later, in Introducing OpenAI o3 and o4-mini | 77.0% | CorroboratedThe same number is printed in 1 other lab document: 1 |
Sources
Reported by OpenAI
- OpenAI o3-miniopenai.com · 31 Jan 2025
Original image - Introducing deep researchopenai.com · 2 Feb 2025
Our capture of the source - Introducing GPT-4.1 in the APIopenai.com · 14 Apr 2025
Our capture of the source - Introducing OpenAI o3 and o4-miniopenai.com · 16 Apr 2025
Our capture of the source
