Autoregressive vision-language-action (VLA) models have recently demonstrated strong capabilities in robotic manipulation. However, their core process of action tokenization often involves a trade-off between reconstruction fidelity and inference efficiency. We introduce FASTer, a unified framework for efficient and generalizable robot learning that integrates a learnable tokenizer with an autoregressive policy built upon it. FASTerVQ encodes action chunks as single-channel images, capturing global spatio-temporal dependencies while maintaining a high compression ratio. FASTerVLA builds on this tokenizer with block-wise autoregressive decoding and a lightweight action expert, achieving both faster inference and higher task performance. Extensive experiments across simulated and real-world benchmarks show that FASTerVQ delivers superior reconstruction quality, high token utilization, and strong cross-task and cross-embodiment generalization, while FASTerVLA further improves overall capability, surpassing previous state-of-the-art VLA models in both inference speed and task performance.
@article{arxiv.2512.04952,
title = {FASTer: Toward Efficient Autoregressive Vision Language Action Modeling via Neural Action Tokenization},
author = {Yicheng Liu and Shiduo Zhang and Zibin Dong and Baijun Ye and Tianyuan Yuan and Xiaopeng Yu and Linqi Yin and Chenhao Lu and Junhao Shi and Luca Jiang-Tao Yu and Liangtao Zheng and Tao Jiang and Jingjing Gong and Xipeng Qiu and Hang Zhao},
journal= {arXiv preprint arXiv:2512.04952},
year = {2025}
}