Qwen 3 0.6B (Thinking): benchmark results
Qwen 3 0.6B evaluated with thinking enabled. Provider: Alibaba. Released 2025-04-29. Access: Open.
Aggregate Rating 1286 ± 12, rank #1899 of 2168 rated models, from 137 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| PreAct-Bench - Daily Life | 54.5 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 62.5 |
| PreAct-Bench - Cybersecurity | 62.6 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 56.2 |
| PreAct-Bench | 58.5 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 43.8 |
| PreAct-Bench - Politics | 58.8 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 43.8 |
| GeoBenchLLM - PPNL Multi-Goal | 7 | Optimal route ratio (%) | 42.9 |
| GeoBenchLLM - PPNL Single-Goal | 60 | Optimal route ratio (%) | 42.9 |
| PreAct-Bench - Law and Contract | 56.3 | Prefix Foresight F1 with the full prefix (PFF(100), binary F | 37.5 |
| AA Humanity's Last Exam | 5.64 | Accuracy (%) | 33.6 |
| GeoBenchLLM - GridRoute | 41 | Optimal route ratio (%) | 28.6 |
| GeoBenchLLM - NY-POI | 33 | Accuracy (%) | 28.6 |
| GeoBenchLLM - SpartUN | 38 | Accuracy (%) | 28.6 |
| GeoBenchLLM - StepGame | 22 | Accuracy (%) | 28.6 |
Interactive version: theaggregate.ai/model?slug=qwen-3-0-6b-thinking · How It Works · Data refreshed daily, snapshot 2026-10-20.