Visual Logical Reasoning on VisuLogic (Accuracy, Token Count, EoT)
27.8AccuracySketchThinker-R1-7B
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SketchThinker-R1-7BMethod Category=Reinforcement-Learning-based, Backbone=Qwen2.5-VL-7B-Instruct2026.01 | 27.8 | 56.3 | 0.494 | |
| Vanilla-R1Method Category=Direct Inference, Backbone=Qwen2.5-VL-7B-Instruct2026.01 | 27.6 | 240 | 0.115 | |
| VeriThinkerMethod Category=Supervised-fine-tuning-based, Backbone=Qwen2.5-VL-7B-Instruct2026.01 | 27.5 | 107.4 | 0.256 | |
| L1Method Category=Reinforcement-Learning-based, Backbone=Qwen2.5-VL-7B-Instruct2026.01 | 27.2 | 167.4 | 0.162 | |
| C3oTMethod Category=Supervised-fine-tuning-based, Backbone=Qwen2.5-VL-7B-Instruct2026.01 | 27.1 | 134.6 | 0.201 | |
| ThinkPruneMethod Category=Reinforcement-Learning-based, Backbone=Qwen2.5-VL-7B-Instruct2026.01 | 26.9 | 183.2 | 0.147 | |
| Chain-of-DraftMethod Category=Prompt-based, Backbone=Qwen2.5-VL-7B-Instruct2026.01 | 26.5 | 94.6 | 0.28 | |
| Constrained CoTMethod Category=Prompt-based, Backbone=Qwen2.5-VL-7B-Instruct2026.01 | 26.4 | 71.5 | 0.369 |