GPT-5.1 (Non-reasoning) — benchmark results

Provider: OpenAI. Released 2025-11-12. Access: API.

Unified ELO 1675 ± 20, rank #372 of 1839 rated models, from 41 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AA Omniscience - Health30.8Accuracy (%)80.9
AA Omniscience - Law24.9Accuracy (%)80
AA Omniscience - Business26.5Accuracy (%)78.7
AA-Omniscience Accuracy29.07Accuracy (%)76.3
AA Omniscience - Software Engineering (SWE) - Julia28Accuracy (%)74.5
AA Omniscience - Humanities & Social Sciences28.3Accuracy (%)74.3
AA Omniscience - Software Engineering (SWE) - PHP40Accuracy (%)73.9
AA Omniscience - Software Engineering (SWE) - Java26Accuracy (%)73.6
AA Omniscience - Software Engineering (SWE) - Swift52Accuracy (%)73.5
AA Omniscience - Software Engineering (SWE) - Python33Accuracy (%)72.8
AA Omniscience - Software Engineering (SWE) - R24Accuracy (%)71
AA Omniscience - Software Engineering (SWE) - Go28Accuracy (%)70.3

Interactive version: theaggregate.ai/model?slug=gpt-5-1-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-08-05.