GPT-5 (Low) — benchmark results
GPT-5 evaluated at the low reasoning-effort setting. Provider: OpenAI. Released 2025-08-07. Access: API.
Unified ELO 1807 ± 23, rank #136 of 1776 rated models, from 63 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Natural Intelligence | 67.2 | NatInt Score | 97.3 |
| AA MATH-500 | 98.73 | Accuracy (%) | 96 |
| AA MMLU-Pro | 85.98 | Accuracy (%) | 93.8 |
| GAIA2 - Ambiguity | 39.6 | Score (%) | 93.3 |
| GAIA2 - Search | 64.2 | Score (%) | 93.3 |
| Aider polyglot coding leaderboard | 81.3 | Pass rate (%) | 93.2 |
| LLM Chess (Saplin) | 858.2 | ELO | 92.1 |
| AA Omniscience - Health | 38.6 | Accuracy (%) | 91.7 |
| UGI - Writing | 48.31 | Writing Score | 88.2 |
| AA Omniscience - Law | 32.1 | Accuracy (%) | 87.5 |
| AA LiveCodeBench | 76.3 | Pass@1 (%) | 87 |
| AA Omniscience - Software Engineering (SWE) - Julia | 40 | Accuracy (%) | 86.8 |
Interactive version: theaggregate.ai/model?slug=gpt-5-low · How the rankings work · Data refreshed daily, snapshot 2026-07-22.