GPT-5.4 Mini (Medium) — benchmark results

GPT-5.4 Mini evaluated at the medium reasoning-effort setting. Provider: OpenAI. Released 2026-03-17. Access: API.

Unified ELO 1662 ± 12, rank #339 of 1776 rated models, from 320 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
EuroEval Slovene NLU - SSJ500K NER83.51Named entity recognition Score (%)100
MageBench S21617Rating100
EuroEval Swedish NLU - SUC382.2Named entity recognition Score (%)99.8
EuroEval Latvian NLU - Latvian Twitter Sentiment55.98Sentiment classification Score (%)99.6
EuroEval Serbian NLU67.51NLU Average Score (%)99.5
EuroEval Icelandic NLU - Hotter and Colder Sentiment61.64Sentiment classification Score (%)99.4
EuroEval Faroese68.52Average Score (%)99.2
EuroEval Bulgarian NLU - BG NER Bsnlp92.21Named entity recognition Score (%)99.1
EuroEval Dutch NLU - DBRD93.85Sentiment classification Score (%)99.1
EuroEval Icelandic NLU - MIM-GOLD NER83.64Named entity recognition Score (%)99
EuroEval Slovene NLU63.05NLU Average Score (%)99
EuroEval Spanish NLU62.2NLU Average Score (%)99

Interactive version: theaggregate.ai/model?slug=gpt-5-4-mini-medium · How the rankings work · Data refreshed daily, snapshot 2026-07-22.