Qwen 3.8 Max (Max): benchmark results
Provider: Alibaba. Released 2026-08-03. Access: API.
Aggregate Rating 1828 ± 25, rank #105 of 2170 rated models, from 18 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SuperCLUE General (July 2026) - Agentic Task Planning | 90.94 | Score | 94.1 |
| SuperCLUE General (July 2026) - Precise Instruction Following | 43.81 | Score | 94.1 |
| SuperCLUE General (July 2026) - Overall | 71.48 | Score | 88.2 |
| SuperCLUE General (July 2026) - Hallucination Control | 86.08 | Score | 79.4 |
| SuperCLUE General (July 2026) - Math Reasoning | 77.19 | Score | 76.5 |
| ASI-Bench (Claude Code) | 38.87 | Scientific Score (0-100) | 75 |
| ASI-Bench (Claude Code) - B1 Full Guidance | 59.14 | Scientific Score (0-100) | 75 |
| ASI-Bench (Claude Code) - B4 Goal + Distractors | 34.46 | Scientific Score (0-100) | 75 |
| SuperCLUE General (July 2026) - Science Reasoning | 71.93 | Score | 64.7 |
| Bug Hunt Bench - VS Code Extension | 12.3 | Planted Bugs Fixed (out of 45) | 64.5 |
| Humanity's Sixth Sense | 33 | Pass@1 (%), all rubric criteria; three attempts/task | 64.3 |
| Bug Hunt Bench | 25.7 | Planted Bugs Fixed (out of 105) | 61.2 |
Interactive version: theaggregate.ai/model?slug=qwen-3-8-max-max · How It Works · Data refreshed daily, snapshot 2026-11-05.