Which Pretraining Paradigm Better Serves Spatial Intelligence? An Empirical Comparison of Vision-Language and Video Generation Models Paper • 2605.28132 • Published May 27 • 25
