Qwen 3 0.6B (Non-reasoning): benchmark results
Qwen 3 0.6B evaluated with thinking disabled. Provider: Alibaba. Released 2025-04-29. Access: Open.
Aggregate Rating 1209 ± 12, rank #2016 of 2170 rated models, from 340 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| PreAct-Bench - Daily Life | 66.7 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 87.5 |
| EuroEval Catalan Summarization - Dacsa CA | 37.49 | chrF3++ (%) | 83.1 |
| PreAct-Bench | 66.7 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 71.9 |
| PreAct-Bench - Cybersecurity | 66.7 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 71.9 |
| PreAct-Bench - Law and Contract | 66.7 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 68.8 |
| EuroEval German Summarization - Mlsum DE | 33.56 | chrF3++ (%) | 66.2 |
| PreAct-Bench - Politics | 66.7 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 65.6 |
| EuroEval Ukrainian Summarization - LR SUM UK | 28.12 | chrF3++ (%) | 63.4 |
| EuroEval Italian Summarization - Ilpost SUM | 36.81 | chrF3++ (%) | 59.3 |
| EuroEval Serbian Summarization - LR SUM SR | 27.29 | chrF3++ (%) | 58.5 |
| EuroEval Spanish Summarization - Mlsum ES | 25.56 | chrF3++ (%) | 58.5 |
| EuroEval Dutch Summarization - Wiki Lingua NL | 32.71 | chrF3++ (%) | 58 |
Interactive version: theaggregate.ai/model?slug=qwen-3-0-6b-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-05.