General Multimodal Capabilities on MMMU-Pro
68.1AccuracyQwen3-VL-32B-Thinking
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-VL-32B-ThinkingParameters=32B, Thinking Mode=true2026.01 | 68.1 | |
| Qwen3-VL-8B-ThinkingParameters=8B, Thinking Mode=true2026.01 | 60.4 | |
| EvoCUA-32BParameters=32B2026.01 | 59.16 | |
| EvoCUA-8BParameters=8B2026.01 | 53.3 | |
| EvoCUA-OpenCUA-72BBase Model=OpenCUA-72B2026.01 | 46.51 | |
| OpenCUA-72BParameters=72B2026.01 | 43.04 |