Xiaomi OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866910202361544704 |
|---|---|
| author | Lu, Jinghui Guan, Jiayi Huang, Zhijian Li, Jinlong Li, Guang Kong, Lingdong Li, Yingyan Wang, Han Xu, Shaoqing Luo, Yuechen Li, Fang Dang, Chenxu Wang, Junli Xu, Tao Wu, Jing Wu, Jianhua Hao, Xiaoshuai Zhang, Wen Jiang, Tianyi Zhang, Lingfeng Zhou, Lei Tang, Yingbo Wang, Jie Gao, Yinfeng Bu, Xizhou Tian, Haochen Qiu, Yihang Jia, Feiyang Liu, Lin Ge, Yigu Li, Hanbing Shen, Yuannan Cui, Jianwei Xie, Hongwei Wang, Bing Sun, Haiyang Zhao, Jingwei Huang, Jiahui Liu, Pei Zhu, Zeyu Jiang, Yuncheng Guo, Zibin Gong, Chuhong Leng, Hanchao Ma, Kun Wang, Naiyan Chen, Guang Yang, Kuiyuan Ye, Hangjun Chen, Long |
| author_facet | Lu, Jinghui Guan, Jiayi Huang, Zhijian Li, Jinlong Li, Guang Kong, Lingdong Li, Yingyan Wang, Han Xu, Shaoqing Luo, Yuechen Li, Fang Dang, Chenxu Wang, Junli Xu, Tao Wu, Jing Wu, Jianhua Hao, Xiaoshuai Zhang, Wen Jiang, Tianyi Zhang, Lingfeng Zhou, Lei Tang, Yingbo Wang, Jie Gao, Yinfeng Bu, Xizhou Tian, Haochen Qiu, Yihang Jia, Feiyang Liu, Lin Ge, Yigu Li, Hanbing Shen, Yuannan Cui, Jianwei Xie, Hongwei Wang, Bing Sun, Haiyang Zhao, Jingwei Huang, Jiahui Liu, Pei Zhu, Zeyu Jiang, Yuncheng Guo, Zibin Gong, Chuhong Leng, Hanchao Ma, Kun Wang, Naiyan Chen, Guang Yang, Kuiyuan Ye, Hangjun Chen, Long |
| contents | Chain-of-Thought (CoT) reasoning has become a powerful driver of trajectory prediction in VLA-based autonomous driving, yet its autoregressive nature imposes a latency cost that is prohibitive for real-time deployment. Latent CoT methods attempt to close this gap by compressing reasoning into continuous hidden states, but consistently fall short of their explicit counterparts. We suggest that this is due to purely linguistic latent representations compressing a symbolic abstraction of the world, rather than the causal dynamics that actually govern driving. Thus, we present OneVL (One-step latent reasoning and planning with Vision-Language explanations), a unified VLA and World Model framework that routes reasoning through compact latent tokens supervised by dual auxiliary decoders. Alongside a language decoder that reconstructs text CoT, we introduce a visual world model decoder that predicts future-frame tokens, forcing the latent space to internalize the causal dynamics of road geometry, agent motion, and environmental change. A three-stage training pipeline progressively aligns these latents with trajectory, language, and visual objectives, ensuring stable joint optimization. In inference, the auxiliary decoders are discarded, and all latent tokens are prefilled in a single parallel pass, matching the speed of answer-only prediction. Across four benchmarks, OneVL becomes the first latent CoT method to surpass explicit CoT, delivering superior accuracy at answer-only latency. These results show that with world model supervision, latent CoT produces more generalizable representations than verbose token-by-token reasoning. Code has been open-sourced to the community. Project Page: https://xiaomi-embodied-intelligence.github.io/OneVL |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_18486 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Xiaomi OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation Lu, Jinghui Guan, Jiayi Huang, Zhijian Li, Jinlong Li, Guang Kong, Lingdong Li, Yingyan Wang, Han Xu, Shaoqing Luo, Yuechen Li, Fang Dang, Chenxu Wang, Junli Xu, Tao Wu, Jing Wu, Jianhua Hao, Xiaoshuai Zhang, Wen Jiang, Tianyi Zhang, Lingfeng Zhou, Lei Tang, Yingbo Wang, Jie Gao, Yinfeng Bu, Xizhou Tian, Haochen Qiu, Yihang Jia, Feiyang Liu, Lin Ge, Yigu Li, Hanbing Shen, Yuannan Cui, Jianwei Xie, Hongwei Wang, Bing Sun, Haiyang Zhao, Jingwei Huang, Jiahui Liu, Pei Zhu, Zeyu Jiang, Yuncheng Guo, Zibin Gong, Chuhong Leng, Hanchao Ma, Kun Wang, Naiyan Chen, Guang Yang, Kuiyuan Ye, Hangjun Chen, Long Computer Vision and Pattern Recognition Computation and Language Robotics Chain-of-Thought (CoT) reasoning has become a powerful driver of trajectory prediction in VLA-based autonomous driving, yet its autoregressive nature imposes a latency cost that is prohibitive for real-time deployment. Latent CoT methods attempt to close this gap by compressing reasoning into continuous hidden states, but consistently fall short of their explicit counterparts. We suggest that this is due to purely linguistic latent representations compressing a symbolic abstraction of the world, rather than the causal dynamics that actually govern driving. Thus, we present OneVL (One-step latent reasoning and planning with Vision-Language explanations), a unified VLA and World Model framework that routes reasoning through compact latent tokens supervised by dual auxiliary decoders. Alongside a language decoder that reconstructs text CoT, we introduce a visual world model decoder that predicts future-frame tokens, forcing the latent space to internalize the causal dynamics of road geometry, agent motion, and environmental change. A three-stage training pipeline progressively aligns these latents with trajectory, language, and visual objectives, ensuring stable joint optimization. In inference, the auxiliary decoders are discarded, and all latent tokens are prefilled in a single parallel pass, matching the speed of answer-only prediction. Across four benchmarks, OneVL becomes the first latent CoT method to surpass explicit CoT, delivering superior accuracy at answer-only latency. These results show that with world model supervision, latent CoT produces more generalizable representations than verbose token-by-token reasoning. Code has been open-sourced to the community. Project Page: https://xiaomi-embodied-intelligence.github.io/OneVL |
| title | Xiaomi OneVL: One-Step Latent Reasoning and Planning with Vision-Language Explanation |
| topic | Computer Vision and Pattern Recognition Computation and Language Robotics |
| url | https://arxiv.org/abs/2604.18486 |