Saved in:
Bibliographic Details
Main Authors: Tur, Yalcin, Naghiyev, Jalal, Fang, Haoquan, Tsai, Wei-Chuan, Duan, Jiafei, Fox, Dieter, Krishna, Ranjay
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.07845
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911431698415616
author Tur, Yalcin
Naghiyev, Jalal
Fang, Haoquan
Tsai, Wei-Chuan
Duan, Jiafei
Fox, Dieter
Krishna, Ranjay
author_facet Tur, Yalcin
Naghiyev, Jalal
Fang, Haoquan
Tsai, Wei-Chuan
Duan, Jiafei
Fox, Dieter
Krishna, Ranjay
contents Current Vision-Language-Action (VLA) models rely on fixed computational depth, expending the same amount of compute on simple adjustments and complex multi-step manipulation. While Chain-of-Thought (CoT) prompting enables variable computation, it scales memory linearly and is ill-suited for continuous action spaces. We introduce Recurrent-Depth VLA (RD-VLA), an architecture that achieves computational adaptivity via latent iterative refinement rather than explicit token generation. RD-VLA employs a recurrent, weight-tied action head that supports arbitrary inference depth with a constant memory footprint. The model is trained using truncated backpropagation through time (TBPTT) to efficiently supervise the refinement process. At inference, RD-VLA dynamically allocates compute using an adaptive stopping criterion based on latent convergence. Experiments on challenging manipulation tasks show that recurrent depth is critical: tasks that fail entirely (0 percent success) with single-iteration inference exceed 90 percent success with four iterations, while simpler tasks saturate rapidly. RD-VLA provides a scalable path to test-time compute in robotics, replacing token-based reasoning with latent reasoning to achieve constant memory usage and up to 80x inference speedup over prior reasoning-based VLA models. Project page: https://rd-vla.github.io/
format Preprint
id arxiv_https___arxiv_org_abs_2602_07845
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Recurrent-Depth VLA: Implicit Test-Time Compute Scaling of Vision-Language-Action Models via Latent Iterative Reasoning
Tur, Yalcin
Naghiyev, Jalal
Fang, Haoquan
Tsai, Wei-Chuan
Duan, Jiafei
Fox, Dieter
Krishna, Ranjay
Robotics
Current Vision-Language-Action (VLA) models rely on fixed computational depth, expending the same amount of compute on simple adjustments and complex multi-step manipulation. While Chain-of-Thought (CoT) prompting enables variable computation, it scales memory linearly and is ill-suited for continuous action spaces. We introduce Recurrent-Depth VLA (RD-VLA), an architecture that achieves computational adaptivity via latent iterative refinement rather than explicit token generation. RD-VLA employs a recurrent, weight-tied action head that supports arbitrary inference depth with a constant memory footprint. The model is trained using truncated backpropagation through time (TBPTT) to efficiently supervise the refinement process. At inference, RD-VLA dynamically allocates compute using an adaptive stopping criterion based on latent convergence. Experiments on challenging manipulation tasks show that recurrent depth is critical: tasks that fail entirely (0 percent success) with single-iteration inference exceed 90 percent success with four iterations, while simpler tasks saturate rapidly. RD-VLA provides a scalable path to test-time compute in robotics, replacing token-based reasoning with latent reasoning to achieve constant memory usage and up to 80x inference speedup over prior reasoning-based VLA models. Project page: https://rd-vla.github.io/
title Recurrent-Depth VLA: Implicit Test-Time Compute Scaling of Vision-Language-Action Models via Latent Iterative Reasoning
topic Robotics
url https://arxiv.org/abs/2602.07845