GPT-5 Pro: benchmark results
OpenAI's higher-compute GPT-5 variant for the most demanding GPT-5 workloads. Provider: OpenAI. Released 2025-08-07. Access: API.
Aggregate Rating 1781 ± 13, rank #81 of 1656 rated models, from 78 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| DriveCombo - L1 Atomic Rules | 86.91 | Accuracy (%) on DriveCombo Level 1 (one atomic traffic rule) | 100 |
| DriveCombo - L2 Static Rule Integration | 83.66 | Accuracy (%) on DriveCombo Level 2 (two coexisting static ru | 100 |
| DriveCombo - L3 Dynamic Interactions | 72.06 | Accuracy (%) on DriveCombo Level 3 (two coexisting rules abo | 100 |
| DriveCombo - L4 Static and Dynamic Rules | 69.82 | Accuracy (%) on DriveCombo Level 4 (a static and a dynamic r | 100 |
| DriveCombo - L5 Rule Conflicts | 44.19 | Accuracy (%) on DriveCombo Level 5 (two conflicting rules re | 100 |
| DriveCombo-Text - L1 Atomic Rules | 89.23 | Accuracy (%) on DriveCombo Level 1 (one atomic traffic rule) | 100 |
| DriveCombo-Text - L2 Static Rule Integration | 86.96 | Accuracy (%) on DriveCombo Level 2 (two coexisting static ru | 100 |
| DriveCombo-Text - L4 Static and Dynamic Rules | 75.78 | Accuracy (%) on DriveCombo Level 4 (a static and a dynamic r | 100 |
| DriveCombo-Text - L5 Rule Conflicts | 47.42 | Accuracy (%) on DriveCombo Level 5 (two conflicting rules re | 100 |
| TrackingAI IQ Test (Mensa Norway) | 97.14 | Mensa Norway Score (%) | 96.1 |
| IKP - T3 | 99.45 | Accuracy (%) | 95.4 |
| M3-Bench - L2 | 89 | BTA Score | 95 |
Interactive version: theaggregate.ai/model?slug=gpt-5-pro · How It Works · Data refreshed daily, snapshot 2026-10-20.