GPT-6 Sol (High): benchmark results
Provider: OpenAI. Released 2026-09-22. Access: API.
Aggregate Rating 1864 ± 12, rank #67 of 2178 rated models, from 90 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA GDP.pdf - Real Estate - Criterion Pass Rate | 79.21 | Criterion Pass Rate (%) | 98 |
| AA Long Context Reasoning | 83.67 | Accuracy (%) | 97.1 |
| MLCR-AA - Accuracy | 94.44 | Accuracy (%) | 97.1 |
| ParseBench - Tables | 92.65 | GTRM Score (GriTS + TableRecordMatch) | 95.8 |
| AA GDP.pdf - STEM/Research - Criterion Pass Rate | 79.19 | Criterion Pass Rate (%) | 94.3 |
| Next.js Agent Evals (Codex) - Success Rate | 100 | Evals passed, pass@4 (%) | 93.8 |
| AutomationBench-AA - Sales - Raw Objective Completion | 86.84 | Raw Objective Completion (%) | 93.5 |
| AA CritPt | 25.43 | Accuracy (%) | 92 |
| AA Humanity's Last Exam | 44.11 | Accuracy (%) | 92 |
| AutomationBench-AA - Operations - Raw Objective Completion | 95.08 | Raw Objective Completion (%) | 90.5 |
| AutomationBench | 31.2 | Strict Task Pass Rate (%) | 90.1 |
| AutomationBench-AA - HR - Raw Objective Completion | 75.57 | Raw Objective Completion (%) | 89.9 |
Interactive version: theaggregate.ai/model?slug=gpt-6-sol-high · How It Works · Data refreshed daily, snapshot 2026-11-07.