Latent Reasoning VLA: Latent Thinking and Prediction for Vision-Language-Action Models

Shuanghao Bai, Jing Lyu, Wanqi Zhou, Zhe Li, Dakai Wang, Lei Xing, Xiaoguang Zhao, Pengwei Wang, Zhongyuan Wang, Cheng Chi, Badong Chen, Shanghang Zhang
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:5409-5426, 2026.

Abstract

Vision-Language-Action (VLA) models benefit from Chain-of-Thought (CoT) reasoning, but existing approaches incur high inference overhead and rely on discrete reasoning representations that mismatch continuous perception and control. We propose Latent Reasoning VLA (LaRA-VLA), a unified VLA framework that internalizes multi-modal CoT reasoning into continuous latent representations for embodied action. LaRA-VLA performs unified reasoning and prediction in latent space, eliminating explicit CoT generation at inference time and enabling efficient, action-oriented control. To realize latent embodied reasoning, we introduce a curriculum-based training paradigm that progressively transitions from explicit textual and visual CoT supervision to latent reasoning, and finally adapts latent reasoning dynamics to condition action generation. We construct two structured CoT datasets, LIBERO-LaRA and Bridge-LaRA, and evaluate LaRA-VLA across simulation benchmarks and long-horizon real-robot manipulation tasks. Experimental results show that LaRA-VLA outperforms existing state-of-the-art VLA methods while achieving up to a 90% reduction in inference latency compared to explicit CoT-based VLA approaches, highlighting latent reasoning as an effective and efficient paradigm for real-time embodied control.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-bai26h, title = {Latent Reasoning {VLA}: Latent Thinking and Prediction for Vision-Language-Action Models}, author = {Bai, Shuanghao and Lyu, Jing and Zhou, Wanqi and Li, Zhe and Wang, Dakai and Xing, Lei and Zhao, Xiaoguang and Wang, Pengwei and Wang, Zhongyuan and Chi, Cheng and Chen, Badong and Zhang, Shanghang}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {5409--5426}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/bai26h/bai26h.pdf}, url = {https://proceedings.mlr.press/v306/bai26h.html}, abstract = {Vision-Language-Action (VLA) models benefit from Chain-of-Thought (CoT) reasoning, but existing approaches incur high inference overhead and rely on discrete reasoning representations that mismatch continuous perception and control. We propose Latent Reasoning VLA (LaRA-VLA), a unified VLA framework that internalizes multi-modal CoT reasoning into continuous latent representations for embodied action. LaRA-VLA performs unified reasoning and prediction in latent space, eliminating explicit CoT generation at inference time and enabling efficient, action-oriented control. To realize latent embodied reasoning, we introduce a curriculum-based training paradigm that progressively transitions from explicit textual and visual CoT supervision to latent reasoning, and finally adapts latent reasoning dynamics to condition action generation. We construct two structured CoT datasets, LIBERO-LaRA and Bridge-LaRA, and evaluate LaRA-VLA across simulation benchmarks and long-horizon real-robot manipulation tasks. Experimental results show that LaRA-VLA outperforms existing state-of-the-art VLA methods while achieving up to a 90% reduction in inference latency compared to explicit CoT-based VLA approaches, highlighting latent reasoning as an effective and efficient paradigm for real-time embodied control.} }
Endnote
%0 Conference Paper %T Latent Reasoning VLA: Latent Thinking and Prediction for Vision-Language-Action Models %A Shuanghao Bai %A Jing Lyu %A Wanqi Zhou %A Zhe Li %A Dakai Wang %A Lei Xing %A Xiaoguang Zhao %A Pengwei Wang %A Zhongyuan Wang %A Cheng Chi %A Badong Chen %A Shanghang Zhang %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-bai26h %I PMLR %P 5409--5426 %U https://proceedings.mlr.press/v306/bai26h.html %V 306 %X Vision-Language-Action (VLA) models benefit from Chain-of-Thought (CoT) reasoning, but existing approaches incur high inference overhead and rely on discrete reasoning representations that mismatch continuous perception and control. We propose Latent Reasoning VLA (LaRA-VLA), a unified VLA framework that internalizes multi-modal CoT reasoning into continuous latent representations for embodied action. LaRA-VLA performs unified reasoning and prediction in latent space, eliminating explicit CoT generation at inference time and enabling efficient, action-oriented control. To realize latent embodied reasoning, we introduce a curriculum-based training paradigm that progressively transitions from explicit textual and visual CoT supervision to latent reasoning, and finally adapts latent reasoning dynamics to condition action generation. We construct two structured CoT datasets, LIBERO-LaRA and Bridge-LaRA, and evaluate LaRA-VLA across simulation benchmarks and long-horizon real-robot manipulation tasks. Experimental results show that LaRA-VLA outperforms existing state-of-the-art VLA methods while achieving up to a 90% reduction in inference latency compared to explicit CoT-based VLA approaches, highlighting latent reasoning as an effective and efficient paradigm for real-time embodied control.
APA
Bai, S., Lyu, J., Zhou, W., Li, Z., Wang, D., Xing, L., Zhao, X., Wang, P., Wang, Z., Chi, C., Chen, B. & Zhang, S.. (2026). Latent Reasoning VLA: Latent Thinking and Prediction for Vision-Language-Action Models. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:5409-5426 Available from https://proceedings.mlr.press/v306/bai26h.html.

Related Material