Towards More Efficient Decoding for Autoregressive Vision-language-action Models

The practical deployment of autoregressive (AR) Vision-Language-Action (VLA) models is severely constrained by inference speed bottlenecks, particularly in high-frequency and dexterous manipulation tasks. While recent studies have explored Jacobi decoding as a more efficient alternative to traditional autoregressive decoding, its practical benefits are marginal due to the lengthy iterations. To address this problem, we introduce consistency distillation to teach the model to predict multiple correct action tokens in each iteration, thereby reducing the total iterations. While the distillation brings moderate speedup, we identify that certain redundancy iterations remain a critical limitation. To tackle this, we propose an adaptive early-exit decoding strategy that moderately relaxes convergence conditions, which further improves average inference efficiency. Experimental results show that the proposed method achieves more than 4× inference acceleration across different base models while maintaining high task success rates in both simulated and real-world robot tasks. These experiments validate that our approach provides an efficient and general paradigm for accelerating multimodal decision-making in robotics. Our code and videos could be found in the supplementary files.