GPT-5.4 Nano (Medium) — benchmark results

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

Unified ELO 1583 ± 9, rank #520 of 1776 rated models, from 318 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
EuroEval Spanish NLU - Sentiment Headlines ES53.49Sentiment classification Score (%)99.3
EuroEval Portuguese NLU - ScaLA PT61.35Linguistic acceptability Score (%)96.6
EuroEval Latvian NLU - Latvian Twitter Sentiment52.2Sentiment classification Score (%)96.1
EuroEval Spanish NLU58.08NLU Average Score (%)95.4
EuroEval Spanish NLU - ScaLA ES49.61Linguistic acceptability Score (%)95.4
EuroEval Estonian NLU - Grammar ET48.69Linguistic acceptability Score (%)95.2
EuroEval German NLU - ScaLA DE62.04Linguistic acceptability Score (%)95.1
EuroEval Slovak NLU - ScaLA SK60.28Linguistic acceptability Score (%)94.6
EuroEval Ukrainian Knowledge84.07Knowledge Average Score (%)94.6
EuroEval Faroese NLU - FoSent70.97Sentiment classification Score (%)94.2
EuroEval Polish Summarization - PSC25.21Score (%)93.9
EuroEval Bosnian NLU - MMS BS50.56Sentiment classification Score (%)93.4

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