Qwen 3.8 Max (0902) (Max): benchmark results
Provider: Alibaba. Released 2026-09-02. Access: API.
Aggregate Rating 1820 ± 25, rank #120 of 2245 rated models, from 16 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SuperCLUE General (July 2026) - Overall | 72.62 | Score | 100 |
| SuperCLUE General (July 2026) - Math Reasoning | 78.95 | Score | 91.2 |
| SuperCLUE General (July 2026) - Agentic Task Planning | 89.53 | Score | 88.2 |
| SuperCLUE General (July 2026) - Hallucination Control | 86.19 | Score | 88.2 |
| SuperCLUE General (July 2026) - Science Reasoning | 75.44 | Score | 85.3 |
| SPU - Private - Raw | 0.39 | Mean Spearman correlation | 83.3 |
| SPU Private Mean | 0.39 | Spearman correlation | 83.3 |
| SuperCLUE General (July 2026) - Precise Instruction Following | 42.86 | Score | 82.4 |
| ExplorationBench - AlienLogic | 76.2 | Held-out accuracy after exploration (%; Best@3: the highest | 77.8 |
| ExplorationBench - AlienCode | 64.3 | Held-out accuracy after exploration (%; Best@3: the highest | 44.4 |
| ExplorationBench - AlienCode (Open-Book) | 41 | Held-out accuracy with the complete rule set supplied before | 33.3 |
| Terminal-Bench 4.0 (Claude Code) | 26.97 | Resolution Rate (%) | 19.2 |
Interactive version: theaggregate.ai/model?slug=qwen-3-8-max-0902-max · How It Works · Data refreshed daily, snapshot 2026-11-16.