List your company
ResearchFormerly WizWand

Fine-Grained Perception & Understanding on MMVP (test)

60AccuracyQwen2.5-VL-7B + SSL-R1
42.663247.164151.66556.1659Apr 22, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Qwen2.5-VL-7B + SSL-R1Model Scale=7B, Training Strategy=+ SSL-R12026.0460
Qwen2.5-VL-7B + Visual JigsawModel Scale=7B, Training Strategy=+ Visual Jigsaw2026.0459.33
ThinkLite-VL-7BModel Scale=7B, Training Strategy=Representative Reasoning Model2026.0455.33
VL-Cogito-7BModel Scale=7B, Training Strategy=Representative Reasoning Model2026.0455.33
Qwen2.5-VL-3B + SSL-R1Model Scale=3B, Training Strategy=+ SSL-R12026.0454.67
Qwen2.5-VL-7BModel Scale=7B, Training Strategy=Baseline2026.0454.66
LLaVA-Critic-R1-7BModel Scale=7B, Training Strategy=Representative Reasoning Model2026.0453.33
Qwen2.5-VL-3B + Visual JigsawModel Scale=3B, Training Strategy=+ Visual Jigsaw2026.0450.67
Qwen2.5-VL-3BModel Scale=3B, Training Strategy=Baseline2026.0443.33