RoboMonkey: Scaling Test-Time Sampling and Verification for Vision-Language-Action Models

TL;DR

RoboMonkey introduces a test-time scaling framework for vision-language-action models that enhances robustness and accuracy in visuomotor tasks through sampling, perturbation, and verification, with significant improvements demonstrated in diverse environments.

Contribution

The paper presents RoboMonkey, a novel test-time scaling method that leverages sampling, synthetic data, and verification to improve VLA model robustness and generalization.

Findings

25% improvement on out-of-distribution tasks

9% improvement on in-distribution tasks

Synthetic data scaling enhances verification accuracy

Abstract

Vision-Language-Action (VLA) models have demonstrated remarkable capabilities in visuomotor control, yet ensuring their robustness in unstructured real-world environments remains a persistent challenge. In this paper, we investigate test-time scaling through the lens of sampling and verification as means to enhance the robustness and generalization of VLAs. We first demonstrate that the relationship between action error and the number of generated samples follows an exponentiated power law across a range of VLAs, indicating the existence of inference-time scaling laws. Building on these insights, we introduce RoboMonkey, a test-time scaling framework for VLAs. At deployment, RoboMonkey samples a small set of actions from a VLA, applies Gaussian perturbation and majority voting to construct an action proposal distribution, and then uses a Vision Language Model (VLM)-based verifier to select the optimal action. We propose a synthetic data generation pipeline for training such VLM-based action verifiers, and demonstrate that scaling the synthetic dataset consistently improves verification and downstream accuracy. Through extensive simulated and hardware experiments, we show that pairing existing VLAs with RoboMonkey yields significant performance gains, achieving a 25% absolute improvement on out-of-distribution tasks and 9% on in-distribution tasks. Additionally, when adapting to new robot setups, we show that fine-tuning both VLAs and action verifiers yields a 7% performance increase compared to fine-tuning VLAs alone.