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

Spatial Reasoning on SITE

67.5Accuracy

Human

Updated 1mo ago

Evaluation Results

Method	Links
Human 2026.02		67.5
GEMINI-3-PRO-PREVIEW 2026.02		62.2
GPT-5 2026.02		61.8
GEMINI-2.5-PRO 2026.02		57
Gemini-2.5-Pro 2026.03		57
GeoThinker QWEN3VL-8B 2026.02		55.9
GeoThinker QWEN2.5VL-7B 2026.02		54.8
SEED-1.6 2026.02		54.6
ACE-Brain-0-8B 2026.03		53.1
VG-LLM-8B 2026.02		52.6
RoboBrain2.5-8B 2026.03		52.6
Pelican-VL-7B 2026.03		52.3
VeBrain-7B 2026.03		51.4
VG-LLM-8B* 2026.02		50.5
InternVL3.5-8B 2026.03		50.1
VG-LLM-4B 2026.02		49.8
RoboBrain2.0-7B 2026.03		49.2
Vlaser-8B 2026.03		47.5
QWEN3-VL-8B-INSTRUCT 2026.02		45.8
Qwen3-VL-8B-Inst. 2026.03		45.8
MiMo-Embodied-7B 2026.03		44.8
INTERNVL3-8B 2026.02		41.1
InternVL3-8B 2026.03		41.1
VST-7B-SFT 2026.02		39.6
GPT-4o 2026.03		37.8
QWEN2.5-VL-7B-INSTRUCT 2026.02		37.6
MiMo-VL-7B 2026.03		37.6
BAGEL-7B-MOT 2026.02		37
VST-3B-SFT 2026.02		35.8
QWEN3-VL-2B-INSTRUCT 2026.02		35.6
Qwen3-VL-2B-Inst. 2026.03		35.6
QWEN2.5-VL-3B-INSTRUCT 2026.02		33.1
CAMBRIAN-S-7B 2026.02		33
Qwen2.5-VL-7B-Inst. 2026.03		31.4
VLM-3R-7B 2026.02		31.3
INTERNVL3-2B 2026.02		30
CAMBRIAN-S-3B 2026.02		28.3
SPATIALLADDER-3B 2026.02		27.9
Random Choice 2026.02		0