Driving · 2026-10-14

AUGUstus: Autonomous Driving VLA Model with Unified Generation and Understanding

Yuqi Ye, Caigao Jiang, Zhuonan Wang, Congyuan Pei, Chao Qu, Changhao Peng, Junhong Lin, Qize Yu, Zetian Song, Guangzhao Fu, Chongruo Wu, Shiyu Huang, Shuchang Zhou, Honggou Yang, Wei Gao

arXiv:2610.16985 PDF

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

AUGUstus is an autonomous-driving VLA that unifies diffusion-based action generation with autoregressive CoT understanding in one transformer. It uses “AUGUstus-mode” to generate actions first and make CoT supervision training-only, enabling low-latency inference. AUGUstus-GRPO jointly optimizes action and CoT, improving planning performance.

Why it matters

The authors address two VLA issues: VA degeneration that produces unsafe/unexplainable actions under low-latency designs, and autoregressive CoT/action token generation that is too slow for deployment. AUGUstus keeps VLM spatial understanding coupled to action generation while enabling action-only inference, and reports strong results on NAVSIM and nuScenes without a scorer module.

Method

Unifies diffusion-based continuous action generation and autoregressive CoT understanding in one transformer, where understanding is optimized only to serve generation.
Uses AUGUstus-mode: action generation before CoT, so inference can skip CoT and output actions directly for low latency.
Introduces AUGUstus-GRPO, a unified GRPO post-training approach that jointly optimizes continuous action and discrete CoT streams.

Limitation

The paper notes that real-time deployment is difficult for VLA+scorer systems because generating N candidate trajectories multiplies the expensive VLM pass, making it difficult to run in real time on vehicle hardware.

Abstract (from arXiv)

Vision-language-action (VLA) models show strong potential for end-to-end autonomous driving by leveraging world knowledge and understanding to excel in long-tail scenarios. However, current VLAs have two limitations: 1) Degeneration into vision-action (VA) models. For low latency, regression/flow-based methods directly generate actions but underuse VLM understanding, causing unsafe, unexplainable actions. 2) Autoregressive high latency. Autoregressive methods sequentially generate chain-of-thought (CoT) and action tokens, fully leveraging VLM understanding, but high latency hinders deployment. To enable low-latency deployment while using CoT to assist action generation, we propose AUGUstus, an Autonomous-driving VLA model with Unified Generation and Understanding. We unify diffusion-based action generation and autoregressive CoT understanding in one model, where understanding only serves generation. Joint CoT supervision preserves VLM spatial understanding in shared representation for safer, more accurate trajectories. AUGUstus-mode orders action generation before CoT, so CoT supervises reliable actions during training while inference needs only action output, enabling low latency. AUGUstus-GRPO, a unified GRPO algorithm, jointly optimizes action and CoT, improving action performance. Experiments show AUGUstus achieves 94.6 PDMS on NAVSIMv1 and 90.7 EPDMS on NAVSIMv2, SOTA planning without scorer module. On nuScenes, it achieves average L2 error 0.27 m and average collision rate 0.05%, best among compared methods. To validate AUGUstus practicality, we train on in-house data with 512 NVIDIA H800 GPUs for ~2 days, evaluate on an internal 1,000 km dataset, and preliminarily deploy on a real vehicle, achieving 12 Hz operation. Project page: https://augustus-ad.github.io

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First public release: 2026-10-12 · Indexed by Pub-AI: 2026-10-14 UTC

preprint