HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model

Recent advancements in vision-language models (VLMs) for common-sense reasoning have led to the development of vision-language-action (VLA) models, enabling robots to perform generalized manipulation. Although existing autoregressive VLA methods leverage large-scale pretrained knowledge, they disrupt the continuity of actions. Meanwhile, some VLA methods incorporate an additional diffusion head to predict continuous actions, relying solely on VLM-extracted features, which limits their reasoning capabilities. In this paper, we introduce HybridVLA, a unified framework that seamlessly integrates the strengths of both autoregressive and diffusion policies within a single large language model, rather than simply connecting them. To bridge the generation gap, a collaborative training recipe is proposed that injects the diffusion modeling directly into the next-token prediction. With this recipe, we find that these two forms of action prediction not only reinforce each other but also exhibit varying performance across different tasks. Therefore, we design a collaborative action ensemble mechanism that adaptively fuses these two predictions, leading to more robust control. In experiments, HybridVLA outperforms previous state-of-the-art VLA methods across various simulation and real-world tasks, including both single-arm and dual-arm robots, while demonstrating stable manipulation in previously unseen configurations.

@article{liu2025_2503.10631,
  title={ HybridVLA: Collaborative Diffusion and Autoregression in a Unified Vision-Language-Action Model },
  author={ Jiaming Liu and Hao Chen and Pengju An and Zhuoyang Liu and Renrui Zhang and Chenyang Gu and Xiaoqi Li and Ziyu Guo and Sixiang Chen and Mengzhen Liu and Chengkai Hou and Mengdi Zhao and KC alex Zhou and Pheng-Ann Heng and Shanghang Zhang },
  journal={arXiv preprint arXiv:2503.10631},
  year={ 2025 }
}