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

Spatial Reasoning on SpatialScore hard

30.6Accuracy

GPT-4o

Updated 2mo ago

Evaluation Results

Method	Links
GPT-4o 2025.09		30.6	53.2
MM-Eureka 2025.09		27.1	52.52
VLAA-Thinker-7B 2025.09		22.3	54.2
Chain-of-Focus 2025.09		20.6	50.94
AdaVaR-7B 2025.09		20.4	55.82
DeepEyes 2025.09		20.3	53.72
OVR-7B 2025.09		20	55.76
GRIT-3B 2025.09		19.7	45.8
ViGoRL-7B 2025.09		19.5	50.53
VLAA-Thinker-3B 2025.09		19.4	46.7
LMM-R1 2025.09		18.9	48.8
AdaVaR-3B 2025.09		18.9	50.84
ViGoRL-3B 2025.09		18.2	45.7
Orsta-7B 2025.09		16.4	52.04
Qwen2.5-VL-3B 2025.09		16.2	45.74
LLaVA-OV-7B 2025.09		15.6	-
Qwen2.5-VL-7B 2025.09		15.2	50.9
InternVL3-8B 2025.09		12.9	-