Qwen 3.6 Max Preview: benchmark results
Alibaba's preview of the Qwen 3.6 Max flagship tier. Provider: Alibaba. Released 2026-04-20. Access: API.
Aggregate Rating 1773 ± 8, rank #85 of 1656 rated models, from 375 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HierSVA-B - Proof Success | 95.2 | Assertion proof success rate (%, C2-P): share of generated a | 100 |
| HierSVA-B - Vacuity Rate | 0.1 | Vacuous proof rate (%, C2-V): share of generated assertions | 100 |
| IG-Bench - IG-Exam - Lineage Verification | 17.4 | Exact accuracy (%) on the 182 lineage-verification instances | 100 |
| LexRubric - Clarity and Structure | 94.62 | Dimension score rate (%): points earned from expert-written | 100 |
| LexRubric - Legal Accuracy | 69.51 | Dimension score rate (%): points earned from expert-written | 100 |
| LexRubric - Legal Consultation | 73.17 | Legal consultation (473 instances) score rate (%): points ea | 100 |
| LexRubric - Task Compliance | 87.3 | Dimension score rate (%): points earned from expert-written | 100 |
| Nejumi 4 - ALT Average | 91.81 | Score (%) | 100 |
| Nejumi 4 - GLP - Expression | 99.5 | Score (%) | 100 |
| Nejumi 4 - jaster (0-shot) - JSeM | 85 | Exact match (%) | 100 |
| OpenCompass Language - Instruction Following (CompassBench 2604) | 80 | Score (%) | 100 |
| PawBench - QwenPaw | 78.35 | Overall Score (%) | 100 |
Interactive version: theaggregate.ai/model?slug=qwen-3-6-max-preview · How It Works · Data refreshed daily, snapshot 2026-10-20.