GPT-6.2 Sol (Medium): benchmark results
Provider: OpenAI. Released 2026-11-17. Access: API.
Aggregate Rating 1954 ± 13, rank #22 of 2274 rated models, from 121 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Maze-Bench | 96.61 | Mean Score | 100 |
| AA GDP.pdf - Construction - Criterion Pass Rate | 83.53 | Criterion Pass Rate (%) | 97.8 |
| AA GDP.pdf - Finance/Investing - Criterion Pass Rate | 91.5 | Criterion Pass Rate (%) | 96.7 |
| AA GDP.pdf - Engineering - Criterion Pass Rate | 84.99 | Criterion Pass Rate (%) | 96.4 |
| MLCR-AA - Accuracy | 95 | Accuracy (%) | 95.1 |
| DeepsecBench | 44.74 | Recall-weighted F2 score (%) | 94.9 |
| AA GDP.pdf - Insurance - Criterion Pass Rate | 87.57 | Criterion Pass Rate (%) | 94.2 |
| AA Terminal-Bench Science 0.1 - Engineering Sciences | 66.67 | Pass@1 (%) | 93.9 |
| AA GDP.pdf | 29.2 | All Criteria Pass Rate (%) | 93.6 |
| AA Terminal-Bench Science 0.1 - Earth Sciences | 66.67 | Pass@1 (%) | 93.6 |
| AutomationBench-AA - Strict | 41.4 | Strict Task Success (%) | 93.1 |
| AutomationBench-AA | 70.28 | Guardrail-Adjusted Objective Completion (%) | 92.4 |
Interactive version: theaggregate.ai/model?slug=gpt-6-2-sol-medium · How It Works · Data refreshed daily, snapshot 2026-11-18.