Qwen 3 4B 2507 (Thinking): benchmark results

Qwen 3 4B 2507 evaluated with thinking enabled. Provider: Alibaba. Released 2025-07-01. Access: Open.

Unified ELO 1487 ± 1, rank #928 of 1761 rated models, from 369 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
EuroEval Slovak NLU - UNER SK73.36Named entity recognition Score (%)100
EuroEval German NLU - GermEval78.27Named entity recognition Score (%)99.4
EuroEval Dutch NLU - CoNLL NL80.14Named entity recognition Score (%)98.8
EuroEval Italian NLU - MultiNERD IT87.46Named entity recognition Score (%)98.8
EuroEval Spanish NLU60.1NLU Average Score (%)97.3
EuroEval Latvian NLU - Fullstack NER LV73.2Named entity recognition Score (%)97
EuroEval Ukrainian NLU - NER UK79.05Named entity recognition Score (%)97
MERA Code - ruHumanEval78.11pass@1 (%)96.9
EuroEval Norwegian NLU - NorNE NB86.13Named entity recognition Score (%)96.7
MERA - ruHumanEval77.44pass@1 (%)96.6
EuroEval Danish NLU - Dansk68.78Named entity recognition Score (%)96.3
EuroEval French NLU - Eltec73.38Named entity recognition Score (%)96

Interactive version: theaggregate.ai/model?slug=qwen-3-4b-2507-thinking · How It Works · Data refreshed daily, snapshot 2026-09-05.