List your company

Spatial Reasoning on CV-Bench (Component Metrics)

86.5Overall ScoreSpatialStack
49.99659.47368.9578.427Mar 28, 2026Apr 7, 2026Apr 17, 2026Apr 27, 2026May 7, 2026May 17, 2026May 28, 2026

Evaluation results

MethodLinks
SpatialStackzero-shot=true2026.0386.56992.593.790.7————
GASP - Full Model (Lcorr + Ldepth)Backbone=Qwen2.5-VL-7B2026.0579.8————68.288.387.379.2
Qwen2.5-VL-32B2026.0579.7————68.980.886.585.8
LLaVA-OneVision-72B2026.0579.7————70.289.282.579
GASP - Correspondence (Lcorr)Backbone=Qwen2.5-VL-7B2026.0579.4————6888.186.678.6
Qwen2.5-VL-7B-Instruct2026.0576.6————63.787.785.572.7
InternVL2.5-4B2026.0574.1————6879.980.769.2
LLaVA-OneVision-7B2026.0573.2————69.277.981.765
LLaMA-3.2V-11B-CoTChain-of-Thought=true2026.0572.8————59.178.978.878
SpatialRGPTzero-shot=true2026.0372.760.478.98071.3————
Qwen2.5-VL-3B-Instruct2026.0568.5————62.670.37864.7
Spatialbotzero-shot=true2026.036861.473.176.561————
Cambrain-8B2026.0562.2————60.781.75550.5
LLaMA-3.2V-11B2026.0558.2————5955.967.350.5
LLaVA-1.5-13B2026.0558.2————56.157.266.853.3
SpaceLLaVA-13B2026.0558.2————56.157.266.853.3
Kimi-VL-3B-Thinking2026.0557.5————60.579.143.544
SpaceQwen2.5VL-3B2026.0551.4————62.245.445.450