StreamingVLA: Streaming Vision-Language-Action Model with Action Flow Matching and Adaptive Early Observation

2026-03-30

AI Summary

This work addresses the high latency and frequent pauses inherent in traditional vision-language-action (VLA) models, which stem from their serial execution of perception, decision-making, and action stages—limiting their suitability for responsive, fluent operation on edge devices. To overcome this, the authors propose an asynchronous parallel streaming architecture that replaces conventional chunk-wise action denoising with action stream matching and integrates an adaptive early observation mechanism guided by action saliency. This design enables overlapping of observation, decision, and execution phases, effectively masking pipeline delays. Without compromising task performance, the approach reduces end-to-end latency by 2.4× and decreases execution stalls by 6.5×, substantially enhancing system responsiveness and interaction smoothness.

Technology Category

• Computer Vision: Language and Vision
• Machine Learning: Hardware-aware ML
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Vision-language-action (VLA) models have demonstrated exceptional performance in natural language-driven perception and control. However, the high computational cost of VLA models poses significant efficiency challenges, particularly for resource-constrained edge platforms in real-world deployments. However, since different stages of VLA (observation, action generation and execution) must proceed sequentially, and wait for the completion of the preceding stage, the system suffers from frequent halting and high latency. To address this, We conduct a systematic analysis to identify the challenges for fast and fluent generation, and propose enabling VLAs with the ability to asynchronously parallelize across VLA stages in a "streaming" manner. First, we eliminate the reliance on action chunking and adopt action flow matching, which learns the trajectory of action flows rather than denoising chunk-wise actions. It overlaps the latency of action generation and execution. Second, we design an action saliency-aware adaptive observation mechanism, thereby overlapping the latency of execution and observation. Without sacrificing performance, StreamingVLA achieves substantial speedup and improves the fluency of execution. It achieves a 2.4 $\times$ latency speedup and reduces execution halting by 6.5 $\times$.

Problem

• Vision-Language-Action
• latency
• real-time execution
• edge computing
• action generation

Innovation

• StreamingVLA
• Action Flow Matching
• Adaptive Observation
• Asynchronous Parallelization
• Vision-Language-Action