Hy-Embodied-RxBrain is an embodied cognition foundation model that jointly reasons over language and visual imagination to produce embodied plans. It merges high‑level task reasoning with physical state modeling in a single planning sequence, unlike vision‑language models that focus on scene understanding or generative world models that only predict future visual states. The model is introduced by Liang, Chen, Huang, Guo, and Zhu (2026) and is available on arXiv (https://arxiv.org/abs/2607.14187).