TwinVLA: Data-Efficient Bimanual Manipulation with Twin Single-Arm Vision-Language-Action Models

TL;DR

TwinVLA introduces a modular approach that composes two pretrained single-arm vision-language-action models to efficiently perform bimanual manipulation tasks, reducing data needs and improving performance.

Contribution

It presents a novel modular framework that combines pretrained single-arm models for effective bimanual manipulation without extensive bimanual training data.

Findings

Outperforms monolithic models on bimanual tasks in real and simulated environments.

Requires no additional bimanual pretraining data.

Narrowing the gap to state-of-the-art models with extensive proprietary data.

Abstract

Vision-language-action models (VLAs) trained on large-scale robotic datasets have demonstrated strong performance on manipulation tasks, including bimanual tasks. However, because most public datasets focus on single-arm demonstrations, adapting VLAs for bimanual tasks typically requires substantial additional bimanual data and fine-tuning. To address this challenge, we introduce TwinVLA, a modular framework that composes two copies of a pretrained single-arm VLA into a coordinated bimanual VLA. Unlike monolithic cross-embodiment models trained on mixtures of single-arm and bimanual data, TwinVLA improves both data efficiency and performance by composing pretrained single-arm policies. Across diverse bimanual tasks in real-world and simulation settings, TwinVLA outperforms a comparably-sized monolithic RDT-1B model without requiring any bimanual pretraining. Furthermore, it narrows the gap to state-of-the-art model $\pi_0$, which relies on extensive proprietary bimanual data and compute cost. These results establish our modular composition approach as a data-efficient and scalable path toward high-performance bimanual manipulation, leveraging public single-arm data.