Qwen 3 Max (2025-09-23): benchmark results
September 23, 2025 Qwen 3 Max snapshot, tracked when sources report the dated API model. Provider: Alibaba. Released 2025-09-23. Access: API.
Aggregate Rating 1676 ± 15, rank #255 of 1656 rated models, from 72 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| MATH Level 5 | 97.13 | Accuracy (%) | 94.4 |
| ReLE - Reasoning - BBH | 83.7 | Accuracy (%) | 83.8 |
| Wolfram LLM Benchmarking Project | 57.2 | Correct Functionality (%) | 79.2 |
| SuperCLUE General (2025 Annual) - Code Generation | 47.23 | Score | 78.8 |
| Chatbot Arena (Text - Multi-Turn) | 1447 | Arena Score | 77 |
| ADRD-Bench - Caregiving QA | 91.95 | Exact-match accuracy (0-1, shown times 100) on 149 caregivin | 76.5 |
| Chatbot Arena (Text - Russian) | 1430 | Arena Score | 72.5 |
| Chatbot Arena (Text - German) | 1438 | Arena Score | 72 |
| Chatbot Arena (Text - Writing, Literature & Language) | 1405 | Arena Score | 71.7 |
| ReLE - Finance | 81.4 | Accuracy (%) | 71.6 |
| Chatbot Arena (Text - Hard Prompts) | 1446 | Arena Score | 71.4 |
| Chatbot Arena (Text - Creative Writing) | 1393 | Arena Score | 71.2 |
Interactive version: theaggregate.ai/model?slug=qwen-3-max-2025-09-23 · How It Works · Data refreshed daily, snapshot 2026-10-20.