Multimodal Mathematical Reasoning on WeMath v1.0 (test)
31.8AccuracyVOLD
Evaluation Results
| Method | Links | |
|---|---|---|
| VOLDImages in FT=false2025.10 | 31.8 | |
| VLAA-Thinker 3BImages in FT=true2025.10 | 31.5 | |
| XReasoner-3B (repl.)Images in FT=false2025.10 | 30.6 | |
| VLM-R1 3B-MathImages in FT=true2025.10 | 30 | |
| Qwen2.5-VL-3BImages in FT=-2025.10 | 22.9 |