Multimodal Reasoning on MMMU-Pro Vision v1.0 (test)
32AccuracyVOLD
Evaluation Results
| Method | Links | |
|---|---|---|
| VOLDImages in FT=false2025.10 | 32 | |
| XReasoner-3B (repl.)Images in FT=false2025.10 | 31 | |
| VLM-R1 3B-MathImages in FT=true2025.10 | 28.6 | |
| Qwen2.5-VL-3BImages in FT=-2025.10 | 27.1 | |
| VLAA-Thinker 3BImages in FT=true2025.10 | 24.6 |