Multimodal Reasoning on VisuLogic (Accuracy)
29.8AccuracyDUPL
Evaluation results
| Method | Links | |
|---|---|---|
| DUPLBackbone=Qwen2.5-VL-7B, Training Algorithm=DUPL2025.10 | 29.8 | |
| GRPOBackbone=Qwen2.5-VL-7B, Training Algorithm=GRPO2025.10 | 28.5 | |
| Qwen2.5-VL-7BBackbone=Qwen2.5-VL-7B2025.10 | 26 |