A Survey on Vision-Language-Action Models: An Action Tokenization Perspective

TL;DR

This survey reviews vision-language-action models, focusing on how they use various types of action tokens to ground and generate executable actions, aiming to unify understanding and guide future research in this evolving field.

Contribution

It categorizes and interprets existing VLA models through action tokenization, providing a comprehensive framework and identifying promising directions for future development.

Findings

Unified framework for VLA models via action tokens

Categorization of action token types and their strengths/limitations

Guidance for future research in VLA model development

Abstract

The remarkable advancements of vision and language foundation models in multimodal understanding, reasoning, and generation has sparked growing efforts to extend such intelligence to the physical world, fueling the flourishing of vision-language-action (VLA) models. Despite seemingly diverse approaches, we observe that current VLA models can be unified under a single framework: vision and language inputs are processed by a series of VLA modules, producing a chain of \textit{action tokens} that progressively encode more grounded and actionable information, ultimately generating executable actions. We further determine that the primary design choice distinguishing VLA models lies in how action tokens are formulated, which can be categorized into language description, code, affordance, trajectory, goal state, latent representation, raw action, and reasoning. However, there remains a lack of comprehensive understanding regarding action tokens, significantly impeding effective VLA development and obscuring future directions. Therefore, this survey aims to categorize and interpret existing VLA research through the lens of action tokenization, distill the strengths and limitations of each token type, and identify areas for improvement. Through this systematic review and analysis, we offer a synthesized outlook on the broader evolution of VLA models, highlight underexplored yet promising directions, and contribute guidance for future research, hoping to bring the field closer to general-purpose intelligence.