Kun Ma2, Guang Chen2, Hangjun Ye2, Wenyu Liu1, Xinggang Wang1✉
1 Huazhong University of Science and Technology 2 Xiaomi EV
(*) Equal contribution. (†) Project leader. (✉)Corresponding Author.
[2026/3/25] Weights of DriveLaW-Video and DriveLaW-Act have been released.🚀
[2026/3/25] Codes of DriveLaW has been released.🚀
[2026/2/21] Our paper has been accepted at CVPR 2026. 🎉
[2025/12/30] ArXiv paper release. Models/Code are coming soon. Please stay tuned! ☕️
- Release Paper
- Release inference & training codes
- Release model weights
World models have become crucial for autonomous driving, as they learn how scenarios evolve over time to address the long-tail challenges of the real world. However, current approaches relegate world models to limited roles: they operate within ostensibly unified architectures that still keep world prediction and motion planning as decoupled processes. To bridge this gap, we propose DriveLaW, a novel paradigm that unifies video generation and motion planning. By directly injecting the latent representation from its video generator into the planner, DriveLaW ensures inherent consistency between high-fidelity future generation and reliable trajectory planning. Specifically, DriveLaW consists of two core components: DriveLaW-Video, our powerful world model that generates high-fidelity forecasting with expressive latent representations, and DriveLaW-Act, a diffusion planner that generates consistent and reliable trajectories from the latent of DriveLaW-Video, with both components optimized by a three-stage progressive training strategy. The power of our unified paradigm is demonstrated by new state-of-the-art results across both tasks. DriveLaW not only advances video prediction significantly, surpassing best-performing work by 33.3% in FID and 1.8% in FVD, but also achieves a new record on the NAVSIM planning benchmark.
The codebase is organized into two main components:
- DriveLaW-Video: Video world model
- DriveLaW-Act: Diffusion-based planner that consumes video latents from DriveLaW-Video
Basic installation for DriveLaW with python==3.10 :
cd DriveLaW-Act
pip install -e .
cd DriveLaW-Video/Train
pip install -e .For a video-only world model :
For planning / NavSim evaluation :
Download weights:
If you have any questions, please contact Tianze Xia via email (xiatianze@hust.edu.cn).
DriveLaW is inspired by the following outstanding contributions to the open-source community: NAVSIM, LTX-Video , ReCogDrive, Diffusers, Genie Envisioner, Epona.
If you find DriveLaW is useful in your research or applications, please consider giving us a star 🌟 and citing it by the following BibTeX entry.
@article{xia2025drivelaw,
title={DriveLaW: Unifying Planning and Video Generation in a Latent Driving World},
author={Xia, Tianze and Li, Yongkang and Zhou, Lijun and Yao, Jingfeng and Xiong, Kaixin and Sun, Haiyang and Wang, Bing and Ma, Kun and Ye, Hangjun and Liu, Wenyu and others},
journal={arXiv preprint arXiv:2512.23421},
year={2025}
}