ChatGPT 4o: benchmark results
Provider: OpenAI. Released 2024-08-15. Access: API.
Aggregate Rating 1660 ± 31, rank #296 of 1635 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AgentDrive | 82.5 | Accuracy (%) | 100 |
| AgentDrive - Hybrid | 72.5 | Accuracy (%) | 100 |
| EIBench (complex) | 39.27 | Recall | 100 |
| VERHallu - Counterfactual QA | 68.5 | Accuracy (%; 967 seven-option questions about an event absen | 100 |
| VERHallu - Subevent QA | 42.5 | Accuracy (%; 258 seven-option questions on which event occur | 100 |
| AgentDrive - Policy | 100 | Accuracy (%) | 99 |
| VERHallu - Causal QA | 46.5 | Accuracy (%; 497 seven-option questions on why an event happ | 94.1 |
| AgentDrive - Physics | 55 | Accuracy (%) | 90 |
| AgentDrive - Comparative | 90 | Accuracy (%) | 75 |
| OmniCoT | 31.58 | Overall accuracy (%), question-weighted over the See, Locate | 75 |
| OmniCoT - Move | 33.65 | Accuracy (%) on Move questions (translational and rotation-t | 75 |
| Roboflow Vision Evals - Visual Understanding | 70.15 | Pass Rate (%) | 73 |
Interactive version: theaggregate.ai/model?slug=chatgpt-4o · How It Works · Data refreshed daily, snapshot 2026-11-05.