List your company

Multimodal Reasoning on GQA (test)

67.4AccuracyOurs
31.41640.75850.159.442Apr 12, 2026

Evaluation results

MethodLinks
OursModel=Qwen2-VL 7B2026.0467.49.20.82
IVT-LRModel=Qwen2-VL 7B2026.0465.810.10.68
Chain-of-FocusModel=Qwen2-VL 7B2026.0461.8128.63.01
CCoTModel=Qwen2-VL 7B2026.0451.276.47.21
SCAFFOLDModel=Qwen2-VL 7B2026.0448.772.86.72
OursModel=Chameleon 7B2026.0439.49.21.21
IVT-LRModel=Chameleon 7B2026.0438.110.10.98
Chain-of-FocusModel=Chameleon 7B2026.0434.6360.42.98
CCoTModel=Chameleon 7B2026.0433.1150.65.31
SCAFFOLDModel=Chameleon 7B2026.0432.81564.17