Latent Reasoning VLA: Latent Thinking and Prediction for Vision-Language-Action Models

TL;DR

LaRA-VLA introduces a continuous latent reasoning framework for vision-language-action models, reducing inference latency and improving performance in embodied tasks by internalizing multi-modal reasoning.

Contribution

It proposes a unified latent reasoning approach that replaces explicit chain-of-thought generation, with a curriculum-based training paradigm for efficient real-time embodied control.

Findings

Outperforms state-of-the-art VLA methods on benchmarks

Reduces inference latency by up to 90%

Effective in long-horizon real-robot manipulation tasks

Abstract

Vision-Language-Action (VLA) models benefit from chain-of-thought (CoT) reasoning, but existing approaches incur high inference overhead and rely on discrete reasoning representations that mismatch continuous perception and control. We propose Latent Reasoning VLA (LaRA-VLA), a unified VLA framework that internalizes multi-modal CoT reasoning into continuous latent representations for embodied action. LaRA-VLA performs unified reasoning and prediction in latent space, eliminating explicit CoT generation at inference time and enabling efficient, action-oriented control. To realize latent embodied reasoning, we introduce a curriculum-based training paradigm that progressively transitions from explicit textual and visual CoT supervision to latent reasoning, and finally adapts latent reasoning dynamics to condition action generation. We construct two structured CoT datasets and evaluate LaRA-VLA on both simulation benchmarks and long-horizon real-robot manipulation tasks. Experimental results show that LaRA-VLA consistently outperforms state-of-the-art VLA methods while reducing inference latency by up to 90\% compared to explicit CoT-based approaches, demonstrating latent reasoning as an effective and efficient paradigm for real-time embodied control. Project Page: https://loveju1y.github.io/Latent-Reasoning-VLA/