Ant spirit wave open source embodiment world model LingBot-VA.

On January 30th, Ant Lingbo Technology announced the open sourcing of the LingBot-VA embodied world model. According to the introduction, this model proposes for the first time a self-regressive video-action world modeling framework, integrating the ability of large-scale video generation models with robot control. The model not only generates the "next world state" simultaneously but also directly infers and outputs the corresponding action sequence, allowing robots to "deduce and act" like humans.