TY  - RPRT
TI  - Which Pretraining Paradigm Better Serves Spatial Intelligence? An Empirical Comparison of Vision-Language and Video Generation Models
AU  - Haozhan Shen
AU  - Tiancheng Zhao
AU  - Kangjia Zhao
AU  - Jianwei Yin
PY  - 2026
UR  - https://arxiv.org/abs/2605.28132
ID  - 2605.28132
ER  -