Share your thoughts, 1 month free Claude Pro on usSee more

Visual Reasoning on CLEVR VL-ICL Bench

38.9Average Accuracy

LLaVA-OneVision-7B

Updated 1mo ago

Evaluation Results

Method	Links
LLaVA-OneVision-7B 2025.06		38.9
LLaVA-OneVision-72B 2025.06		37.2
LLaVA-OneVision-7B 2025.06		34.9
LAQ-7B + MAPD 2025.06		31.4
LAQ-7B + In-ContextPD 2025.06		30.9
Qwen2.5-VL-7B-Instruct 2025.06		29.1
LLaVA-Next-7B 2025.06		28.6
Qwen2-VL-7B-Instruct 2025.06		27.9
LAQ-7B + In-ContextPD 2025.06		24.1
Qwen2.5-VL-7B-Instruct 2025.06		22
LLaVA-Next-7B 2025.06		21.1
LLaVA v1.5-7B 2025.06		16.9
LAQ-7B + MAPD 2025.06		12.3
LLaVA v1.5-7B 2025.06		10.9