GPT-6 Sol (xHigh): benchmark results
Provider: OpenAI. Released 2026-09-22. Access: API.
Aggregate Rating 1890 ± 14, rank #49 of 2178 rated models, from 73 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| DeepsecBench | 40.75 | Recall-weighted F2 score (%) | 100 |
| AA GDP.pdf - Real Estate - Criterion Pass Rate | 79.31 | Criterion Pass Rate (%) | 98.4 |
| AutomationBench-AA - Sales - Raw Objective Completion | 90.07 | Raw Objective Completion (%) | 98.4 |
| QuantArena - Iterative - Overall | 51.14 | Score (%) | 97.3 |
| AutomationBench-AA - Operations - Raw Objective Completion | 96.51 | Raw Objective Completion (%) | 96.8 |
| AA CritPt | 28 | Accuracy (%) | 94.9 |
| MLCR-AA - Accuracy | 93.33 | Accuracy (%) | 94.9 |
| DataBench | 61.3 | Score (%) | 94.4 |
| AA GDP.pdf - STEM/Research - Criterion Pass Rate | 79.01 | Criterion Pass Rate (%) | 93.5 |
| AA Humanity's Last Exam | 46.29 | Accuracy (%) | 93.4 |
| MLCR-AA | 32.22 | Overall Score (%) | 93.4 |
| AutomationBench | 33.2 | Strict Task Pass Rate (%) | 92.8 |
Interactive version: theaggregate.ai/model?slug=gpt-6-sol-xhigh · How It Works · Data refreshed daily, snapshot 2026-11-07.