SOTA Spatial Reasoning on MindCube and PapersWithCode

94.5Accuracy

Human

Updated 1mo ago

Evaluation Results

Method	Links
Human 2026.02		94.5
Human 2026.02		94.5
SenseNova-SI (InternVL3-8B) 2026.02		85.6
GeoThinker QWEN2.5VL-7B 2026.02		83.6
GeoThinker QWEN3VL-8B 2026.02		83
ACE-Brain-0-8B 2026.03		82.1
GEMINI-3-PRO-PREVIEW 2026.02		70.8
Gemini-3 Pro 2026.02		70.8
Grok-4 2026.02		63.5
SSR-2D 2026.02		63.5
GEMINI-2.5-PRO 2026.02		57.6
Gemini-2.5-Pro 2026.01		57.6
Gemini-2.5-Pro 2026.03		57.6
GPT-5 2026.02		56.3
GPT-5 2026.02		56.3
GPT-4o 2026.01		53.4
MindCube-3B-RawQA-SFT 2026.02		51.7
SEED-1.6 2026.02		48.7
Seed-1.6 2026.02		48.7
GPT-4o 2026.03		46.1
SPATIALLADDER-3B 2026.02		43.4
SpatialLadder-3B 2026.01		43.4
INTERNVL3-8B 2026.02		41.5
InternVL3-8B 2026.01		41.5
InternVL3-8B 2026.03		41.5
VLM-3R-7B 2026.02		40
VLM-3R-7B 2026.02		40
VST-7B-SFT 2026.02		39.7
VST-7B-SFT 2026.01		39.7
CAMBRIAN-S-7B 2026.02		39.6
SmoothOp-7B 2026.01		39.6
Cambrian-S-7B 2026.02		39.6
SmoothOp-3B 2026.01		39.4
SpaceR-7B 2026.01		37.9
QWEN2.5-VL-3B-INSTRUCT 2026.02		37.6
Qwen2.5-VL-3B 2026.01		37.6
INTERNVL3-2B 2026.02		37.5
VG-LLM-4B 2026.02		36.9
Claude-4-Sonnet 2026.03		36.6
VG-LLM-8B* 2026.02		36.1
QWEN2.5-VL-7B-INSTRUCT 2026.02		36
Qwen2.5-VL-7B 2026.01		36
Qwen2.5-VL-7B-Inst. 2026.03		36
VST-3B-SFT 2026.02		35.9
VST-3B-SFT 2026.01		35.9
ViLaSR-7B 2026.01		35.1
InternVL3.5-8B 2026.03		35.1
BAGEL-7B-MOT 2026.02		34.7
LLava-OneVision-7B 2026.01		34.7
Bagel-7B-MoT 2026.02		34.7
Vlaser-8B 2026.03		34.6
Qwen3-VL-2B-Inst. 2026.03		34.5
Random Choice 2026.02		33
Random Choice 2026.02		33
VG-LLM-8B 2026.02		32.7
CAMBRIAN-S-3B 2026.02		32.5
MiMo-Embodied-7B 2026.03		32.3
VST-7B-SFT 2026.02		32
QWEN3-VL-2B-INSTRUCT 2026.02		31.4
RoboBrain2.0-7B 2026.03		31.2
Pelican-VL-7B 2026.03		31
ViLaSR-7B 2026.02		30.2
VeBrain-7B 2026.03		30.1
QWEN3-VL-8B-INSTRUCT 2026.02		29.8
Qwen3-VL-8B-Inst. 2026.03		29.4
RoboBrain2.5-8B 2026.03		28.1
SpatialLadder-3B 2026.02		27.4
SpaceR-7B 2026.02		27.4
Spatial-MLLM-4B 2026.02		26.1