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

Spatial Relationship Reasoning on SPAR-Bench

59.9Accuracy (Avg)

GAMSI_S1+S2

Updated 8d ago

Evaluation Results

Method	Links
GAMSI_S1+S2 2026.05		59.9	-	-	-
GAMSI_S1 2026.05		49.6	-	-	-
SenseNova-SI-InternVL3-8B 2026.05		48.9	-	-	-
VST-7B-SFT 2026.05		47.2	-	-	-
SenseNova-SI-Qwen3-VL-8B 2026.05		41.1	-	-	-
Qwen3-VL-8B-Instruct 2026.05		39.9	-	-	-
Bagel-7B-MoT 2026.05		39.1	-	-	-
VST-3B-SFT 2026.05		38.1	-	-	-
Cambrian-S-7B 2026.05		38	-	-	-
SpaceR-7B 2026.05		37.3	-	-	-
ViLaSR-7B 2026.05		37.3	-	-	-
GPT-4o 2026.05		36.4	29.3	24.9	45.1
InternVL3-8B 2026.05		35.8	-	-	-
Spatial-MLLM-4B 2026.05		35.3	-	-	-
Qwen2.5-VL-7B-Instruct 2026.05		33.8	-	-	-
Qwen3-VL-2B-Instruct 2026.05		33.3	-	-	-
SpatialLadder-3B 2026.05		32.9	-	-	-
Cambrian-S-3B 2026.05		32.9	-	-	-
LLaVA-OneVision-7B 2026.05		31.2	21.8	26.1	40.1
InternVL-2.5-4B 2026.05		30.6	25.7	29.8	35.2
Qwen2.5-VL-7B 2026.05		30.2	17.5	29.5	41.8
Qwen2.5-VL-3B-Instruct 2026.05		30	-	-	-
Kimi-VL-A3B 2026.05		29.4	26.4	18.8	36.2
CaMo-3B 2026.05		27.6	13	31.3	42.6
InternVL3-2B 2026.05		27.2	-	-	-
Qwen2.5-VL-3B 2026.05		24.6	15.3	26.4	32.2