Qwen vision-language model for visual reasoning, documents, and agent tasks
Benchmarks
| Benchmark | Category | Score | Variant | Provenance | Source | Observed |
|---|---|---|---|---|---|---|
| AA Intelligence Index | intelligence | 33.8 | — | independent | aa | Jul 24, 2026 |
| 29.3 | non-reasoning | independent | aa | Jul 24, 2026 | ||
| LMArena Text | preference | 1429 | — | crowd | lmarena | Jul 21, 2026 |
| LMArena Vision | preference | 1241 | — | crowd | lmarena | Jul 21, 2026 |
| LMArena WebDev | preference | 1357 | — | crowd | lmarena | Jul 21, 2026 |
| Median output speed | speed | 78 tok/s | — | independent | aa | Jul 24, 2026 |
| 83 tok/s | non-reasoning | independent | aa | Jul 24, 2026 | ||
| Median time to first token | speed | 5.63s | — | independent | aa | Jul 24, 2026 |
| 5.61s | non-reasoning | independent | aa | Jul 24, 2026 |
Includes index data from Artificial Analysis.