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

Spatial Reasoning on SPBench SI

79.1Accuracy

GAMSI_S1+S2

Updated 1mo ago

Evaluation Results

Method	Links
GAMSI_S1+S2 2026.05		79.1
SpatialLadder-3B 2026.05		70
OneCanvas 2026.06		62.8
SenseNova-SI-InternVL3-8B 2026.05		62.5
GAMSI_S1 2026.05		60.6
SenseNova-SI-Qwen3-VL-8B 2026.05		60.4
SpaceMind 2026.06		59.7
SpatialLadder-3B 2026.05		58.6
Cambrian-S-3B 2026.05		57.7
Cambrian-S-7B 2026.05		54.6
3ViewSense-4B-rl-slack 2026.03		54.2
VST-3B-SFT 2026.05		53.7
3ViewSense-4B-sft 2026.03		52.6
VST-7B-SFT 2026.05		52.2
3ViewSense-4B-rl-strict 2026.03		52
Qwen3-VL-8B-Instruct 2026.05		51.7
Qwen3-VL-2B-Instruct 2026.05		51.6
CaMo-3B 2026.05		51.5
InternVL3-8B 2026.05		51.5
VILASR-7B 2026.06		50.2
Gemini-2.0-Flash 2026.05		49
ViLaSR-7B 2026.05		48.9
SpaceR-7B 2026.06		48.6
SpaceR-7B 2026.05		47.9
Qwen2.5-VL-7B-Instruct 2026.05		47.4
Spatial-MLLM-4B 2026.05		45.3
Video-R1 2026.06		44.9
Spatial-MLLM-4B 2026.06		43.7
Qwen2.5-VL-3B 2026.03		43.5
InternVL3-2B 2026.05		41.2
Qwen2.5-VL-3B-Instruct 2026.05		40.9
SpatialLadder-3B 2026.03		40.8
Spatial-MLLM-4B 2026.05		38.1
Bagel-7B-MoT 2026.05		38.1
Qwen3-VL-8B-Instruct 2026.03		37.6
VILASR-7B 2026.05		36.6
Qwen2.5-VL-7B 2026.05		36.3
SpaceR-7B 2026.05		35.7
InternVL-2.5-4B 2026.05		31.8
SpaceQwen2.5-VL-3B-Instruct 2026.03		29.7
InternVL-2.5-8B 2026.05		28.3
InternVL3.5-4B 2026.03		28.1
Video-R1 2026.05		27.7
Qwen3-VL-4B-Instruct 2026.03		27.1
Kimi-VL-A3B 2026.05		25.7
LLaVA-OneVision-7B 2026.05		25.4
GPT-4o 2026.05		24.5
Qwen2.5-VL-3B 2026.05		24.3