Visual Logical Reasoning on VisuLogic (Score)
41.7ScoreBAGEL
Evaluation Results
| Method | Links | |
|---|---|---|
| BAGELSize=14B2026.04 | 41.7 | |
| Tuna-2Size=7B2026.04 | 28.8 | |
| EVE (ours-iter3)Backbone=Qwen3-VL-4B, Iteration=32026.04 | 27.3 | |
| EVE (ours-iter2)Backbone=Qwen3-VL-4B, Iteration=22026.04 | 27.2 | |
| Show-o2Size=7B2026.04 | 26.9 | |
| Ming-UniVisionSize=16B2026.04 | 26.7 | |
| Tuna-RSize=7B2026.04 | 26.2 | |
| EVE (ours-iter1)Backbone=Qwen3-VL-4B, Iteration=12026.04 | 26.1 | |
| VILA-USize=7B2026.04 | 25.4 | |
| HarmonSize=1.5B2026.04 | 25.1 | |
| OneCatSize=9B2026.04 | 24.9 | |
| LLaVA-OV2026.04 | 24.8 | |
| Emu3Size=8B2026.04 | 24.7 | |
| TarSize=7B2026.04 | 24.3 | |
| Qwen3-VL-4B-InstructBackbone=Qwen3-VL-4B, Iteration=Base2026.04 | 24.2 | |
| Janus-ProSize=7B2026.04 | 23.8 | |
| TunaSize=7B2026.04 | 22.4 | |
| JanusFlowSize=1.3B2026.04 | 22 | |
| Qwen2.5-VLSize=7B2026.04 | 20 |