A Survey on Efficient Vision-Language-Action Models

2025-10-27

AI Summary

Vision-Language-Action (VLA) models exhibit strong generalization capabilities but face significant deployment barriers in embodied AI due to prohibitive computational costs and large-scale data requirements. To address this, we propose the “Efficient VLA” research framework—a systematic, unified taxonomy spanning model architecture design, training optimization, and robot-centric data acquisition. Methodologically, we integrate lightweight architectures, model compression techniques, sample- and compute-efficient training strategies, and principled approaches for high-yield robotic data collection and utilization. We comprehensively survey state-of-the-art advances, distill representative application paradigms, and identify key challenges including scalability, out-of-distribution generalization, and data bias. Furthermore, we establish a continuously updated open-source project page. This work provides foundational theoretical insights and practical engineering guidance for developing computationally efficient, deployable embodied AI systems.

Technology Category

• Intelligent Robots: Embodied AI
• Computer Vision: Large Vision Models
• Machine Learning: Efficient ML / Green AI

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Vision-Language-Action models (VLAs) represent a significant frontier in embodied intelligence, aiming to bridge digital knowledge with physical-world interaction. While these models have demonstrated remarkable generalist capabilities, their deployment is severely hampered by the substantial computational and data requirements inherent to their underlying large-scale foundation models. Motivated by the urgent need to address these challenges, this survey presents the first comprehensive review of Efficient Vision-Language-Action models (Efficient VLAs) across the entire data-model-training process. Specifically, we introduce a unified taxonomy to systematically organize the disparate efforts in this domain, categorizing current techniques into three core pillars: (1) Efficient Model Design, focusing on efficient architectures and model compression; (2) Efficient Training, which reduces computational burdens during model learning; and (3) Efficient Data Collection, which addresses the bottlenecks in acquiring and utilizing robotic data. Through a critical review of state-of-the-art methods within this framework, this survey not only establishes a foundational reference for the community but also summarizes representative applications, delineates key challenges, and charts a roadmap for future research. We maintain a continuously updated project page to track our latest developments: https://evla-survey.github.io/

Problem

• Surveying efficient Vision-Language-Action models for embodied intelligence
• Addressing computational and data bottlenecks in VLA deployment
• Organizing efficient techniques across data, model, and training

Innovation

• Efficient model design and compression techniques
• Efficient training methods reduce computational burdens
• Efficient data collection addresses robotic bottlenecks