English

StarVLA-$\alpha$: Reducing Complexity in Vision-Language-Action Systems

Robotics 2026-04-14 v1 Artificial Intelligence Computer Vision and Pattern Recognition

Abstract

Vision-Language-Action (VLA) models have recently emerged as a promising paradigm for building general-purpose robotic agents. However, the VLA landscape remains highly fragmented and complex: as existing approaches vary substantially in architectures, training data, embodiment configurations, and benchmark-specific engineering. In this work, we introduce StarVLA-α\alpha, a simple yet strong baseline designed to study VLA design choices under controlled conditions. StarVLA-α\alpha deliberately minimizes architectural and pipeline complexity to reduce experimental confounders and enable systematic analysis. Specifically, we re-evaluate several key design axes, including action modeling strategies, robot-specific pretraining, and interface engineering. Across unified multi-benchmark training on LIBERO, SimplerEnv, RoboTwin, and RoboCasa, the same simple baseline remains highly competitive, indicating that a strong VLM backbone combined with minimal design is already sufficient to achieve strong performance without relying on additional architectural complexity or engineering tricks. Notably, our single generalist model outperforms π0.5\pi_{0.5} by 20\% on the public real-world RoboChallenge benchmark. We expect StarVLA-α\alpha to serve as a solid starting point for future research in the VLA regime. Code will be released at https://github.com/starVLA/starVLA.

Keywords

Cite

@article{arxiv.2604.11757,
  title  = {StarVLA-$\alpha$: Reducing Complexity in Vision-Language-Action Systems},
  author = {Jinhui Ye and Ning Gao and Senqiao Yang and Jinliang Zheng and Zixuan Wang and Yuxin Chen and Pengguang Chen and Yilun Chen and Shu Liu and Jiaya Jia},
  journal= {arXiv preprint arXiv:2604.11757},
  year   = {2026}
}
R2 v1 2026-07-01T12:06:58.972Z