Paper 2603.19199
FASTER: Rethinking Real-Time Flow VLAs
- Published
- Mar 2026
- Research lab
- Independent
- Citations
- 10
- GitHub
- Not linked
01 In brief
Summary
FASTER is a method to reduce reaction latency in flow-based Vision-Language-Action (VLA) models for real-time robot control.
It introduces a Horizon-Aware Schedule (HAS) that adaptively allocates more sampling steps to near-term actions, enabling the immediate action to be generated in a single step while preserving long-horizon trajectory quality.
This compresses the Time to First Action (TTFA) by up to 10x compared to constant schedules used in models like π0.5 and X-VLA.
FASTER also employs a streaming client-server interface with early stopping to increase inference frequency and reduce expected reaction time.
Real-world experiments, including a dynamic table tennis task, show FASTER outperforms synchronous, naive asynchronous, and Training-time RTC baselines, especially on consumer-grade GPUs (RTX 4060).
The method is plug-and-play, requiring no architectural changes, and integrates into standard fine-tuning pipelines.
On RTX 4090, FASTER reduces TTFA from 80ms to 62.1ms for π0.5 and from 113.7ms to 44.8ms for X-VLA, with corresponding improvements in expected reaction time.
On RTX 4060, gains are larger, with X-VLA TTFA dropping from 399.5ms to 129.2ms.
FASTER also improves task completion scores and reduces completion times in real-robot tasks.
02 From the paper
Abstract
Real-time execution is crucial for deploying Vision-Language-Action (VLA) models in the physical world. Existing asynchronous inference methods primarily optimize trajectory smoothness, but neglect the critical latency in reacting to environmental changes. By rethinking the notion of reaction in action chunking policies, this paper presents a systematic analysis of the factors governing reaction time. We show that reaction time follows a uniform distribution determined jointly by the Time to First Action (TTFA) and the execution horizon. Moreover, we reveal that the standard practice of applying a constant schedule in flow-based VLAs can be inefficient and forces the system to complete all sampling steps before any movement can start, forming the bottleneck in reaction latency. To overcome this issue, we propose Fast Action Sampling for ImmediaTE Reaction (FASTER). By introducing a Horizon-Aware Schedule, FASTER adaptively prioritizes near-term actions during flow sampling, compressing the denoising of the immediate reaction by tenfold (e.g., in $π_{0.5}$ and X-VLA) into a single step, while preserving the quality of long-horizon trajectory. Coupled with a streaming client-server pipeline, FASTER substantially reduces the effective reaction latency on real robots, especially when deployed on consumer-grade GPUs. Real-world experiments, including a highly dynamic table tennis task, prove that FASTER unlocks substantially improved real-time responsiveness for generalist policies, enabling rapid generation of accurate and smooth trajectories.