# StreamVLA: Breaking the Reason-Act Cycle via Completion-State Gating

arXiv: https://arxiv.org/abs/2602.01100 · HTML: https://mlsanity.com/paper/2602.01100/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 1 Feb 2026.

## Built on by (most cited first, 7 of 7)

- [WeaveLA: Event Driven Cross-Subtask Latent Memory Weaving for Repetitive Robot Manipulation](https://mlsanity.com/paper/2606.17463/) (2026)
- [PonderPounce: A Pretrained MLLM as an Episode Context Engine for Robot Control](https://mlsanity.com/paper/2608.24115/) (2026)
- [SkipVLA: Skipping VLA Steps with Classical Planning for Fast Robot Manipulation](https://mlsanity.com/paper/2609.20648/) (2026)
- [Toward Unified Robot Learning: Bridging Representation, Vision-Language-Action, and World Models](https://mlsanity.com/paper/2609.03927/) (2026)
- [SLIP-VLA: Single-Step Latent Imagination for Policy Learning in Vision-Language-Action Models](https://mlsanity.com/paper/2609.33575/) (2026)
- [Staircase Policy: Streaming Inference for World-Action Models with Large Action Chunks](https://mlsanity.com/paper/2609.36471/) (2026)
- [Remember What You Did: Action-History Memory with Dual-Expert Denoising for Long-Horizon Vision-Language-Action Policies](https://mlsanity.com/paper/2609.37307/) (2026)

## Abstract (arXiv)

Long-horizon robotic manipulation requires bridging the gap between high-level planning (System 2) and low-level control (System 1). Current Vision-Language-Action (VLA) models often entangle these processes, performing redundant multimodal reasoning at every timestep, which leads to high latency and goal instability. To address this, we present StreamVLA, a dual-system architecture that unifies textual task decomposition, visual goal imagination, and continuous action generation within a single parameter-efficient backbone. We introduce a "Lock-and-Gated" mechanism to intelligently modulate computation: only when a sub-task transition is detected, the model triggers slow thinking to generate a textual instruction and imagines the specific visual completion state, rather than generic future frames. Crucially, this completion state serves as a time-invariant goal anchor, making the policy robust to execution speed variations. During steady execution, these high-level intents are locked to condition a Flow Matching action head, allowing the model to bypass expensive autoregressive decoding for 72% of timesteps. This hierarchical abstraction ensures sub-goal focus while significantly reducing inference latency. Extensive evaluations demonstrate that StreamVLA achieves state-of-the-art performance, with a 98.5% success rate on the LIBERO benchmark and robust recovery in real-world interference scenarios, achieving a 48% reduction in latency compared to full-reasoning baselines.