DM0: An Embodied-Native Vision-Language-Action Model towards Physical AI
Abstract
Moving beyond the traditional paradigm of adapting internet-pretrained models to physical tasks, we present DM0, an Embodied-Native Vision-Language-Action (VLA) framework designed for Physical AI. Unlike approaches that treat physical grounding as a fine-tuning afterthought, DM0 unifies embodied manipulation and navigation by learning from heterogeneous data sources from the onset. Our methodology follows a comprehensive three-stage pipeline: Pretraining, Mid-Training, and Post-Training. First, we conduct large-scale unified pretraining on the Vision-Language Model (VLM) using diverse corpora--seamlessly integrating web text, autonomous driving scenarios, and embodied interaction logs-to jointly acquire semantic knowledge and physical priors. Subsequently, we build a flow-matching action expert atop the VLM. To reconcile high-level reasoning with low-level control, DM0 employs a hybrid training strategy: for embodied data, gradients from the action expert are not backpropagated to the VLM to preserve generalized representations, while the VLM remains trainable on non-embodied data. Furthermore, we introduce an Embodied Spatial Scaffolding strategy to construct spatial Chain-of-Thought (CoT) reasoning, effectively constraining the action solution space. Experiments on the RoboChallenge benchmark demonstrate that DM0 achieves state-of-the-art performance in both Specialist and Generalist settings on Table30.
Cite
@article{arxiv.2602.14974,
title = {DM0: An Embodied-Native Vision-Language-Action Model towards Physical AI},
author = {En Yu and Haoran Lv and Jianjian Sun and Kangheng Lin and Ruitao Zhang and Yukang Shi and Yuyang Chen and Ze Chen and Ziheng Zhang and Fan Jia and Kaixin Liu and Meng Zhang and Ruitao Hao and Saike Huang and Songhan Xie and Yu Liu and Zhao Wu and Bin Xie and Pengwei Zhang and Qi Yang and Xianchi Deng and Yunfei Wei and Enwen Zhang and Hongyang Peng and Jie Zhao and Kai Liu and Wei Sun and Yajun Wei and Yi Yang and Yunqiao Zhang and Ziwei Yan and Haitao Yang and Hao Liu and Haoqiang Fan and Haowei Zhang and Junwen Huang and Yang Chen and Yunchao Ma and Yunhuan Yang and Zhengyuan Du and Ziming Liu and Jiahui Niu and Yucheng Zhao and Daxin Jiang and Wenbin Tang and Xiangyu Zhang and Zheng Ge and Erjin Zhou and Tiancai Wang},
journal= {arXiv preprint arXiv:2602.14974},
year = {2026}
}
Comments
Authors are listed in alphabetical order. Code is available at https://github.com/Dexmal/dexbotic