O3 (Medium): benchmark results
O3 evaluated at the medium reasoning-effort setting. Provider: OpenAI. Released 2025-04-16. Access: API.
Aggregate Rating 1699 ± 16, rank #360 of 2168 rated models, from 57 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BAZAAR (Lechmazur) | 7.37 | TrueSkill μ | 100 |
| Gapminder AI Worldview | 93.9 | Correct Rate (%) | 100 |
| HAL AssistantBench | 38.81 | Accuracy (%) | 100 |
| HAL ScienceAgentBench | 33.33 | Accuracy (%) | 100 |
| TALES | 58.7 | Score (% of max) | 100 |
| TALES - ALFWorld | 88.3 | Score (% of max) | 100 |
| TurtleAI - Real | 40.2 | Success rate (%) on the 102 TurtleAI-DSReal tasks (XLogoOnli | 100 |
| TALES - TextWorld | 100 | Score (% of max) | 98.7 |
| Step Game (Lechmazur) | 5.32 | TrueSkill μ | 98.6 |
| TALES - Jericho | 15.7 | Score (% of max) | 97.4 |
| TALES - ScienceWorld | 93 | Score (% of max) | 97.4 |
| HAL SciCode | 9.23 | Accuracy (%) | 96.7 |
Interactive version: theaggregate.ai/model?slug=o3-medium · How It Works · Data refreshed daily, snapshot 2026-10-20.