Qwen 3.5 0.8B (Thinking): benchmark results
Provider: Alibaba. Released 2026-03-02. Access: Open.
Aggregate Rating 1216 ± 14, rank #2005 of 2170 rated models, from 67 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Tau2-Bench Telecom | 47.66 | Success Rate (%) | 53 |
| MedLayXPlain | 58.2 | S (self-reported) | 51.6 |
| AA TAU-2 Bench | 47.66 | Accuracy (%) | 51.5 |
| AA CritPt | 0 | Accuracy (%) | 20.9 |
| AA Long Context Reasoning | 9 | Accuracy (%) | 14.4 |
| AA Omniscience - Software Engineering (SWE) - Julia | 0 | Accuracy (%) | 10.8 |
| AA Omniscience - Software Engineering (SWE) - TypeScript | 5.56 | Accuracy (%) | 8.7 |
| Swallow - Post-trained English - AIME | 1.7 | Accuracy (%) | 8.2 |
| Swallow - Post-trained Japanese - GPQA | 26.1 | Accuracy (%) | 8.2 |
| FrameBench - Frame Identification - Japanese | 39.3 | Accuracy (%; Japanese FrameNet candidate frames) | 8 |
| AA Omniscience - Software Engineering (SWE) - Go | 4 | Accuracy (%) | 7.7 |
| Swallow - Post-trained English - MMLU-Pro | 38.8 | Accuracy (%) | 7.5 |
Interactive version: theaggregate.ai/model?slug=qwen-3-5-0-8b-thinking · How It Works · Data refreshed daily, snapshot 2026-11-05.