OpenAI · model
o3-pro benchmark scores
As of 2026-10-07, o3-pro (OpenAI) has measured scores on 2 benchmarks and estimated scores on 26 more.
| Benchmark | Score | Source |
|---|---|---|
| AA-GPQA Diamond | 84.5% | measured |
| AA-HLE | 23.8% | estimated ± 4.0 pp, high confidence |
| Artificial Analysis Intelligence Index | 21.9% | measured |
| CritPt | 1.6% | estimated ± 2.9 pp, high confidence |
| GPQA Diamond | 86.0% | estimated ± 2.4 pp, high confidence |
| HLE | 27.5% | estimated ± 9.9 pp, medium confidence |
| AA-Omniscience Accuracy | 27.9% | estimated ± 7.5 pp, medium confidence |
| BioMysteryBench (human-difficult) | 35.2% | estimated ± 8.1 pp, low confidence |
| MMLU-Pro | 85.2% | estimated ± 3.9 pp, high confidence |
| GPQA | 86.7% | estimated ± 3.2 pp, high confidence |
| ARC-AGI-1 | 60.6% | estimated ± 5.4 pp, medium confidence |
| ARC-AGI-2 | 22.7% | estimated ± 14.2 pp, low confidence |
| Vals GPQA Diamond | 79.3% | estimated ± 7.4 pp, medium confidence |
| Vals MMLU-Pro | 84.2% | estimated ± 2.6 pp, medium confidence |
| HLE w/o tools | 27.5% | estimated ± 6.3 pp, medium confidence |
| C-Eval | 91.9% | estimated ± 1.0 pp, medium confidence |
| MMLU-Redux | 94.6% | estimated ± 2.1 pp, medium confidence |
| SuperGPQA | 69.5% | estimated ± 10.4 pp, low confidence |
| HealthBench Hard | 11.2% | estimated ± 7.7 pp, low confidence |
| MMLU-Pro (Arcee) | 84.3% | estimated ± 5.4 pp, low confidence |
| BioMysteryBench (human-solvable) | 49.1% | estimated ± 1.0 pp, low confidence |
| HealthBench (raw) | 30.9% | estimated ± 6.7 pp, low confidence |
| HealthBench (length-adjusted) | 53.9% | estimated ± 3.7 pp, low confidence |
| HealthBench Professional | 55.1% | estimated ± 4.9 pp, low confidence |
| HealthBench Professional (raw) | 40.8% | estimated ± 5.0 pp, low confidence |
| HLE-Verified | 15.3% | estimated ± 2.0 pp, low confidence |
| LABBench2 | 68.9% | estimated ± 1.8 pp, low confidence |
| MMLU | 96.0% | estimated ± 1.9 pp, low confidence |