# SkillVLA: Tackling Combinatorial Diversity in Dual-Arm Manipulation via Skill Reuse

Source: [SkillVLA: Tackling Combinatorial Diversity in Dual-Arm Manipulation via Skill Reuse](<https://www.sota2.com/research/paper/69a90e122d603e6de3e4085b>)

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

Data updated: 2026-03-06T05:00:05.534Z

Authors: Xuanran Zhai, Zekai Huang, Longyan Wu, Qianyou Zhao, Qiaojun Yu, Jieji Ren, Ce Hao, Harold Soh

Published: 2026-03-04T08:38:27.000Z

## Abstract

Recent progress in vision-language-action (VLA) models has demonstrated strong potential for dual-arm manipulation, enabling complex behaviors and generalization to unseen environments. However, mainstream bimanual VLA formulations largely overlook the critical challenge of combinatorial diversity. Different pairings of single-arm behaviors can induce qualitatively distinct task behaviors, yet existing models do not explicitly account for this structure. We argue that effective bimanual VLAs should support skill reuse - the ability to recombine previously learned single-arm skills across novel left-right pairings - thereby avoiding the need to separately learn every possible combination. Current VLA designs entangle skills across arms, preventing such recomposition and limiting scalability. To address this limitation, we propose SkillVLA, a framework explicitly designed to enable skill reuse in dual-arm manipulation. Extensive experiments demonstrate that SkillVLA substantially improves skill composition, increasing overall success rate from 0% to 51%, and achieves strong performance on cooperative and long-horizon tasks.

## 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.

- [Bimanual Manipulation on Skill Recomposition Tasks Unseen Combinations](<https://www.sota2.com/research/sota/bimanual-manipulation-on-skill-recomposition-tasks-unseen-combinations.md>): Bimanual Manipulation; Skill Recomposition Tasks Unseen Combinations; Success Rate (Cup x Cake): 70.

- [Cooperative Bimanual Manipulation on Cooperative Bimanual Tasks](<https://www.sota2.com/research/sota/cooperative-bimanual-manipulation-on-cooperative-bimanual-tasks.md>): Cooperative Bimanual Manipulation; Cooperative Bimanual Tasks; Shake Success Rate: 25.

- [Robot Manipulation on Skills Learned](<https://www.sota2.com/research/sota/robot-manipulation-on-skills-learned.md>): Robot Manipulation; Skills Learned; Success Rate (Cup): 80.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2603.03836>)

[PDF](<https://arxiv.org/pdf/2603.03836.pdf>)