GPT-5.3: benchmark results

OpenAI's GPT-5.3 Instant (March 2026), a fast everyday-conversation chat model in the GPT-5 series focused on smoother, more natural responses for common ChatGPT use. Provider: OpenAI. Released 2026-03-03. Access: API.

Aggregate Rating 1762 ± 9, rank #104 of 1640 rated models, from 276 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
HelpBench - Data Concerns89Data Concerns topic (data collection, exposure, loss and sha100
HelpBench - Moderation Tools84Moderation Tools topic (moderation and anti-abuse tools such100
MapSatisfyBench - Explicit Completion92.72Explicit-factor completion rate (%): share of the query's st100
PerfReasoning - Default Effort3.2Both-Exact Pass Rate (%)100
PhysAI-Bench52Decision accuracy (%; four-option choice of the next action 100
SciDiagramEdit - AutoFigure-Edit Image Aesthetics45.3UniPercept score (0–100)100
Synthetic Hospital - Evidence Retrieval83.3Precision at 5 (%; ranking chart sections for a diagnosis ag100
Synthetic Hospital - Imaging Indication51.8Concept F1 (%; inferring the clinical question behind a vagu100
Evals for Every Language - Basque (eu)74.45Average Score (%)98.4
Evals for Every Language - Southern Sotho (st)72.24Average Score (%)98.4
Evals for Every Language - Tswana (tn)67.41Average Score (%)98.4
HelpBench87Overall rubric score (%): points for met positive criteria a97.1

Interactive version: theaggregate.ai/model?slug=gpt-5-3 · How It Works · Data refreshed daily, snapshot 2026-11-07.