1 paper
Danze Chen, Zeqing Wang, Ziyue Lin +2
We present H3-World, an efficient framework that turns the 33B MiniMax-H3 video generator into an interactive world model. Our key finding is that, as large video generators become…