GPT-5.1 (Thinking): benchmark results
GPT-5.1 evaluated with thinking enabled. Provider: OpenAI. Released 2025-11-12. Access: API.
Aggregate Rating 1694 ± 14, rank #380 of 2168 rated models, from 52 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| GenomeQA (Binary Choice) - Enhancer and Promoter Identification | 70.6 | Accuracy (%) on the two-option (chance 50) Enhancer and Prom | 100 |
| GenomeQA (Binary Choice) - Splice Site Identification | 54.4 | Accuracy (%) on the two-option (chance 50) Splice Site Ident | 100 |
| OpenAI GPT-5.6 System Card - Image Input - Erotic | 99.9 | not_unsafe rate (%) | 100 |
| SuperGPQA | 66.35 | Accuracy (%) | 97.1 |
| CCTU (Thinking) | 16.5 | Perfect solve rate (%): the share of cases in which the mode | 87.5 |
| ALE-Bench | 1192.15 | Performance (Self-Refine x1) (imported) | 83.3 |
| OpenAI GPT-5.6 System Card - Challenging Prompts - Extremism | 100 | not_unsafe rate (%) | 83.3 |
| GenomeQA (Binary Choice) | 60.77 | Accuracy (%) on the two-option (chance 50) GenomeQA question | 80 |
| GenomeQA (Binary Choice) - Taxonomic Classification | 64.6 | Accuracy (%) on the two-option (chance 50) Taxonomic Classif | 80 |
| GenomeQA (Multiple Choice) - Enhancer and Promoter Identification | 62.6 | Accuracy (%) on the four-option (chance 25) Enhancer and Pro | 80 |
| GenomeQA (Multiple Choice) - Histone Mark Prediction | 36.2 | Accuracy (%) on the four-option (chance 25) Histone Mark Pre | 80 |
| GenomeQA (Multiple Choice) - TFBS Prediction | 50.2 | Accuracy (%) on the four-option (chance 25) TFBS Prediction | 80 |
Interactive version: theaggregate.ai/model?slug=gpt-5-1-thinking · How It Works · Data refreshed daily, snapshot 2026-10-20.