Model card
Qwen3-VL-235B-A22B-Thinking.
QwenLarge language model
Added from Papers with Code MMLU-Pro refresh on 2026-05-19.
§ 02 · Benchmarks
Every benchmark Qwen3-VL-235B-A22B-Thinking has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | DocVQA | Computer Vision · Document Understanding | anls | 96.5% | #2 | — | source ↗ |
| 02 | MMBench | Multimodal · Visual Question Answering | accuracy | 88.8% | #6 | — | source ↗ |
| 03 | MVBench | Multimodal · Video Understanding | accuracy | 75.2% | #6 | — | source ↗ |
| 04 | MMMU | Multimodal · Image-Text-to-Text | accuracy | 80.6% | #8 | — | source ↗ |
| 05 | RealWorldQA | Multimodal · Visual Question Answering | accuracy | 81.3% | #8 | — | source ↗ |
| 06 | AIME 2025 | Reasoning · Mathematical Reasoning | accuracy | 89.7% | #9 | — | source ↗ |
| 07 | MMStar | Multimodal · Image-Text-to-Text | accuracy | 78.7% | #9 | — | source ↗ |
| 08 | Video-MME | Multimodal · Video Understanding | accuracy | 79.0% | #9 | — | source ↗ |
| 09 | OSWorld | Agentic AI · Web & Desktop Agents | success-rate | 38.1% | #10 | — | source ↗ |
| 10 | LiveCodeBench | Computer Code · Code Generation | pass-1 | 70.1% | #17 | — | source ↗ |
| 11 | MMMU-Pro | Multimodal · Visual Question Answering | accuracy | 69.3% | #18 | — | source ↗ |
| 12 | GPQA Diamond | Reasoning · Multi-step Reasoning | accuracy | 77.1% | #35 | — | source ↗ |
| 13 | MMLU-Pro | Reasoning · Commonsense Reasoning | accuracy | 83.8% | #37 | 2025-11-26 | source ↗ |
| 14 | MMLU-Pro | Reasoning · Commonsense Reasoning | accuracy | 83.8% | #37 | — | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Strengths by area
Where Qwen3-VL-235B-A22B-Thinking actually performs.
§ 04 · Papers
1 paper with results for Qwen3-VL-235B-A22B-Thinking.
- 2025-11-26· 13 results
Qwen3-VL Technical Report
§ 05 · Related models
Other Qwen models scored on Codesota.
§ 06 · Sources & freshness
Where these numbers come from.
pwc-dump
13
results
paperswithcode
1
result
0 of 14 rows marked verified.