Multimodal Reasoning on AlgoPuzzleVQA
31.83AccuracyCORA
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CORAModel=Qwen2.5-VL-7B2026.06 | 31.83 | 2.5 | |
| BaselineModel=Qwen2.5-VL-7B2026.06 | 31.28 | 12.54 | |
| CORAModel=Qwen2.5-VL-3B2026.06 | 30.17 | 30.38 | |
| BaselineModel=Qwen2-VL-2B2026.06 | 29.39 | 96.55 | |
| BaselineModel=Qwen2.5-VL-3B2026.06 | 28.78 | 31.54 | |
| CORAModel=Qwen2-VL-2B2026.06 | 28.28 | 99.09 | |
| BaselineModel=Qwen2-VL-7B2026.06 | 26.72 | 16.46 | |
| CORAModel=Qwen2-VL-7B2026.06 | 26.67 | 5.31 |