GPT-5.1 (Low): benchmark results
GPT-5.1 evaluated at the low reasoning-effort setting. Provider: OpenAI. Released 2025-11-12. Access: API.
Unified ELO 1638 ± 1, rank #290 of 1761 rated models, from 24 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Natural Intelligence | 61.99 | NatInt Score | 94.7 |
| UGI - Writing | 50.84 | Writing Score | 88.5 |
| LLM Chess (Saplin) | 614.7 | ELO | 77.6 |
| UGI Leaderboard | 44.15 | UGI Score | 77.6 |
| Design Arena (Data Viz) | 1224 | Elo | 66 |
| Epoch AI - Mystery Game Puzzles | 19 | Score | 57.3 |
| FrontierMath - Tiers 1-3 | 17.3 | Accuracy (%, 290 problems) | 54.5 |
| Chess Puzzles (Epoch AI) | 14 | Accuracy (%) | 50.7 |
| Design Arena (Game Dev) | 1181 | Elo | 50.3 |
| Inverse Turing Bench | 63.02 | Accuracy (self-reported) | 50 |
| Design Arena (Website) | 1187 | Elo | 49.1 |
| OTIS Mock AIME 2024-25 | 63.89 | Accuracy (%) | 48.1 |
Interactive version: theaggregate.ai/model?slug=gpt-5-1-low · How It Works · Data refreshed daily, snapshot 2026-09-05.