O3 Pro: benchmark results
OpenAI's pro-tier o-series reasoning model for difficult multi-step tasks. Provider: OpenAI. Released 2025-06-10. Access: API.
Aggregate Rating 1752 ± 14, rank #116 of 1656 rated models, from 61 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Ducky Bench (Stabby Quack) | 1436 | ELO | 100 |
| Aider polyglot coding leaderboard | 84.9 | Pass rate (%) | 97 |
| IneqMath | 46 | Overall Accuracy (imported) | 96.3 |
| AGI-Eval Community - Instruction Following | 64.63 | Accuracy (%) | 94.3 |
| NYT Connections Older Models | 81.7 | Score (%) | 92 |
| AGI-Eval Community - Interaction (Chinese) | 85.87 | Accuracy (%) | 90.6 |
| SEAL - VISTA | 51.63 | Score | 90.5 |
| AGI-Eval Community - Cognition (English) | 85.56 | Accuracy (%) | 88 |
| Merge-Bench | 46.1 | Equivalent text (self-reported) | 87.5 |
| TrackingAI IQ Test (Mensa Norway) | 94.29 | Mensa Norway Score (%) | 86.8 |
| AGI-Eval Community - Cognition | 83.57 | Accuracy (%) | 86.4 |
| AGI-Eval Community - Interaction | 87.17 | Accuracy (%) | 86.1 |
Interactive version: theaggregate.ai/model?slug=o3-pro · How It Works · Data refreshed daily, snapshot 2026-10-20.