Qwen 3.6 35B A3B (Thinking): benchmark results
Provider: Alibaba. Released 2026-04-16. Access: Open.
Aggregate Rating 1660 ± 13, rank #497 of 2201 rated models, from 122 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| VideoJev Evaluation - MLVU development MCQ | 86.2 | Accuracy (%) | 100 |
| Medmarks - LongHealth Task 2 | 90.75 | Score (%) | 97.2 |
| Medmarks - MedHallu Hard | 62.26 | Score (%) | 95.5 |
| Medmarks - MedHallu Medium | 75.32 | Score (%) | 95.5 |
| Medmarks - SuperGPQA Medicine Hard | 55.45 | Score (%) | 95.5 |
| Medmarks - MedHallu Easy | 82.67 | Score (%) | 94.9 |
| Medmarks - PubMedQA | 80.47 | Score (%) | 94.9 |
| Medmarks - SuperGPQA Medicine Easy | 67.29 | Score (%) | 94.3 |
| Nejumi 4 - Toxicity - Violation Categories | 55.56 | Criteria met (%) | 93.5 |
| K-MetBench | 82.9 | Accuracy (imported) | 93.1 |
| Medmarks - MedCalc-Bench | 78.73 | Score (%) | 92 |
| Nejumi 4 - Toxicity - Social Norms | 98.81 | Criteria met (%) | 91.4 |
Interactive version: theaggregate.ai/model?slug=qwen-3-6-35b-a3b-thinking · How It Works · Data refreshed daily, snapshot 2026-11-12.