GPT-5.1: benchmark results
OpenAI's standard GPT-5.1 model with adaptive reasoning for general-purpose chat, coding, and writing. Provider: OpenAI. Released 2025-11-12. Access: API.
Aggregate Rating 1734 ± 3, rank #135 of 1656 rated models, from 1667 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AEGIS (Academic Image Forensics) | 48.8 | Normalized Forensic Index (%): harmonic mean of the macro-F1 | 100 |
| AEGIS (Academic Image Forensics) - Manipulation Classification | 48.34 | Macro-F1 (%) over three classes, naming a highlighted edited | 100 |
| AGC-Bench - data_narrative | 1.45 | Dataset z-score | 100 |
| BABE (Biology Arena) | 52.31 | Average score (%), the paper's fixed weighting of its strong | 100 |
| BABE (Biology Arena) - Strong Correlation | 51.79 | Score (%) on the strong-correlation triplets, where each que | 100 |
| CARV - Difference (Different Source) | 62.4 | Accuracy (%) on 500 CARV visual analogy tasks (controlled sc | 100 |
| CARV - Intersection (Different Source) | 73 | Accuracy (%) on 500 CARV visual analogy tasks (controlled sc | 100 |
| CARV - Intersection (Same Source) | 75 | Accuracy (%) on 500 CARV visual analogy tasks (controlled sc | 100 |
| CARV - Single-Step | 81.2 | Accuracy (%) on 500 CARV visual analogy tasks (controlled sc | 100 |
| CARV - Union (Different Source) | 51 | Accuracy (%) on 500 CARV visual analogy tasks (controlled sc | 100 |
| CARV - Union (Same Source) | 64.6 | Accuracy (%) on 500 CARV visual analogy tasks (controlled sc | 100 |
| CaReCoS | 48.8 | Accuracy (%) of one-to-two-sentence answers to clinically gr | 100 |
Interactive version: theaggregate.ai/model?slug=gpt-5-1 · How It Works · Data refreshed daily, snapshot 2026-10-20.