# Teaching Tiny VLA Models Where to Look and How to Move

Source: [Teaching Tiny VLA Models Where to Look and How to Move](<https://www.sota2.com/research/paper/6a4c882b85dba69fef8b2ac0>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-07-31T05:00:01.455Z

Authors: Iok Tong Lei, Ying Jie Yap, Wei Huang, Qingchen Xie, Qianzhi Li, Yujie Zhang, Xiaolong Liu, Zhidong Deng

Published: 2026-07-05T08:34:39.000Z

## Abstract

Tiny Vision-Language-Action models are appealing for real-time robotic control, but reducing model scale often weakens two capabilities essential for manipulation: task-conditioned spatial grounding and coherent action generation. We introduce XS-VLA, a lightweight framework that teaches tiny VLA policies "where to look" and "how to move" without increasing deployment-time model cost. For spatial grounding, Coarse-Grained Spatial Distillation uses Qwen3-VL-4B to produce teacher-derived coarse image-plane location labels for task-relevant objects, which are quantized into a spatial vocabulary and distilled into a SmolVLM2-0.25B backbone without human annotations. For action generation, Latent Flow Matching combines a CVAE-style latent variable with flow-based policy learning to organize multimodal demonstrations during training and produce stable action chunks at deployment. On LIBERO, XS-VLA improves average success from 82.8\\% to 90.3\\% over SmolVLA-0.25B and improves LIBERO-Long from 63.0\\% to 89.0\\%. In real-world experiments, we design three tasks covering single-arm placement, precision bimanual stacking, and long-horizon sequential coordination, where XS-VLA improves average task success from 21.7\\% to 65.0\\%. These results show that explicit spatial grounding and latent action-structure learning can make tiny VLA models effective for robotic manipulation.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Robot Manipulation on LIBERO (Success Rates and Completion Time)](<https://www.sota2.com/research/sota/robot-manipulation-on-libero-success-rates-and-completion-time.md>): Robot Manipulation; LIBERO; Spatial Success Rate: 94.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2607.04171>)

[PDF](<https://arxiv.org/pdf/2607.04171.pdf>)
