![[2601.09708] Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2601.09708] Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning
Summary
The paper presents Fast-ThinkAct, a novel framework for efficient Vision-Language-Action reasoning that reduces inference latency while maintaining effective planning and adaptability in dynamic environments.
Why It Matters
Fast-ThinkAct addresses the challenge of high inference latency in Vision-Language-Action tasks, which is crucial for real-time applications in robotics and AI. By improving reasoning efficiency, this framework can enhance the performance of AI systems in complex environments, making it relevant for advancements in AI and robotics.
Key Takeaways
- Fast-ThinkAct reduces inference latency by up to 89.3% compared to existing frameworks.
- The framework employs verbalizable latent reasoning to enhance planning capabilities.
- It demonstrates effective long-horizon planning and few-shot adaptation.
- The proposed method is applicable to diverse embodied manipulation tasks.
- Fast-ThinkAct enables better failure recovery in dynamic environments.
Computer Science > Computer Vision and Pattern Recognition arXiv:2601.09708 (cs) [Submitted on 14 Jan 2026 (v1), last revised 24 Feb 2026 (this version, v2)] Title:Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning Authors:Chi-Pin Huang, Yunze Man, Zhiding Yu, Min-Hung Chen, Jan Kautz, Yu-Chiang Frank Wang, Fu-En Yang View a PDF of the paper titled Fast-ThinkAct: Efficient Vision-Language-Action Reasoning via Verbalizable Latent Planning, by Chi-Pin Huang and 6 other authors View PDF HTML (experimental) Abstract:Vision-Language-Action (VLA) tasks require reasoning over complex visual scenes and executing adaptive actions in dynamic environments. While recent studies on reasoning VLAs show that explicit chain-of-thought (CoT) can improve generalization, they suffer from high inference latency due to lengthy reasoning traces. We propose Fast-ThinkAct, an efficient reasoning framework that achieves compact yet performant planning through verbalizable latent reasoning. Fast-ThinkAct learns to reason efficiently with latent CoTs by distilling from a teacher, driven by a preference-guided objective to align manipulation trajectories that transfers both linguistic and visual planning capabilities for embodied control. This enables reasoning-enhanced policy learning that effectively connects compact reasoning to action execution. Extensive experiments across diverse embodied manipulation and reasoning benchmarks demonstrate that Fast-ThinkAc...
