GPT-5 Mini — benchmark results

OpenAI's smaller, cost-efficient GPT-5 tier. Provider: OpenAI. Released 2025-08-07. Access: API.

Unified ELO 1637 ± 7, rank #394 of 1776 rated models, from 758 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
AMA-Bench - Open-World QA78.81Average Score (%)100
AMA-Bench - Text2SQL83.44Average Score (%)100
AMA-Bench - Web82.95Average Score (%)100
CulturaQA75.09CulturaQA (self-reported)100
EuroEval German NLU - Sb10K64.7Sentiment classification Score (%)100
EuroEval Latvian Common Sense Reasoning97.25Common Sense Reasoning Average Score (%)100
EuroEval Romanian NLU - RoSent98.51Sentiment classification Score (%)100
HardcoreLogic - Unsolvable Puzzles98.52Accuracy (%)100
MMTU - Table Understanding93.25Accuracy (%)100
ProLLM - Summarization98.8Score (%)100
EuroEval Latvian NLU - Latvian Twitter Sentiment53.12Sentiment classification Score (%)98.3
EuroEval Polish Common Sense Reasoning69.61Common Sense Reasoning Average Score (%)98.1

Interactive version: theaggregate.ai/model?slug=gpt-5-mini · How the rankings work · Data refreshed daily, snapshot 2026-07-22.