[Beijing: Significantly improve the real application and deployment capabilities of embodied intelligence models, and the embodied intelligence basic model has overall reached the world's first-class level] On November 11, the "Beijing Embodied Intelligent Robot Industry High-Quality Development Action Plan (2026-2028)" was issued, and the plan proposed to significantly improve the capabilities of the embodied intelligence basic model. Support the layout and technology integration development of multiple technical routes such as visual language action model (VLA), world model (WM), reinforcement learning, etc., and lead the technological innovation of each major route. Support the research on technologies such as multi-source information fusion, physical space understanding, and human emotion recognition, and promote the leap from "perception" to "cognition". "Brain" cognition reaches the level of understanding physical law constraints, active recognition of non-verbal instructions such as human actions and expressions; support research on technologies such as real-time reasoning, cross-ontology migration, and self-learning evolution. , greatly improving the "cerebellum" capabilities of multi-task generalization, high-reliability execution, fast-paced operations, and ultra-precision operations; supporting technologies such as group collaboration, long-term memory, localized lightweight deployment, and model interpretability, significantly improving the real application deployment capabilities of embodied intelligence models, and the embodied intelligence basic model as a whole has reached the world's first-class level.