# Humanoid Loco-Manipulation With Discrete VLA Model

arXiv: https://arxiv.org/abs/2609.35709 · HTML: https://mlsanity.com/paper/2609.35709/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Tactile, force-aware and humanoid VLA, one of 13 directions in Vision-language-action models.
- Builds on 21 of its 24 references that are in the map.
- First posted on arXiv 28 Sep 2026.

## Placed in

- Vision-language-action models / [Tactile, force-aware and humanoid VLA](https://mlsanity.com/vla/tactile-force-aware-and-humanoid-vla/)

## Builds on (most cited first, 21 of 21)

- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [π0: A Vision-Language-Action Flow Model for General Robot Control](https://mlsanity.com/paper/2410.24164/) (2024)
- [π0.5: a Vision-Language-Action Model with Open-World Generalization](https://mlsanity.com/paper/2504.16054/) (2025)
- [Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardware](https://mlsanity.com/paper/2304.13705/) (2023)
- [MaskGIT: Masked Generative Image Transformer](https://mlsanity.com/paper/2202.04200/) (2022)
- [FAST: Efficient Action Tokenization for Vision-Language-Action Models](https://mlsanity.com/paper/2501.09747/) (2025)
- [World Action Models are Zero-shot Policies](https://mlsanity.com/paper/2602.15922/) (2026)
- [EgoDex: Learning Dexterous Manipulation from Large-Scale Egocentric Video](https://mlsanity.com/paper/2505.11709/) (2025)
- [Knowledge Insulating Vision-Language-Action Models: Train Fast, Run Fast, Generalize Better](https://mlsanity.com/paper/2505.23705/) (2025)
- [SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Control](https://mlsanity.com/paper/2511.07820/) (2025)
- [Discrete Diffusion VLA: Bringing Discrete Diffusion to Action Decoding in Vision-Language-Action Policies](https://mlsanity.com/paper/2508.20072/) (2025)
- [WholeBodyVLA: Towards Unified Latent VLA for Whole-Body Loco-Manipulation Control](https://mlsanity.com/paper/2512.11047/) (2025)
- [Psi0: An Open Foundation Model Towards Universal Humanoid Loco-Manipulation](https://mlsanity.com/paper/2603.12263/) (2026)
- [Humanoid Everyday: A Comprehensive Robotic Dataset for Open-World Humanoid Manipulation](https://mlsanity.com/paper/2510.08807/) (2025)
- [FASTer: Toward Efficient Autoregressive Vision Language Action Modeling via Neural Action Tokenization](https://mlsanity.com/paper/2512.04952/) (2025)
- [G0.5: One Autoregressive Stream for Robot Reasoning and Action](https://mlsanity.com/paper/2608.11739/) (2026)
- [DiffusionVL: Translating Any Autoregressive Models into Diffusion Vision Language Models](https://mlsanity.com/paper/2512.15713/) (2025)
- [HoloMotion-1 Technical Report](https://mlsanity.com/paper/2605.15336/) (2026)
- [SIMPLE: Simulation-Based Policy Learning and Evaluation for Humanoid Loco-manipulation](https://mlsanity.com/paper/2606.08278/) (2026)

## Abstract (arXiv)

Vision-language-action (VLA) models using discrete action tokens have proven effective for controling robotic arms on manipulation tasks. For a humanoid, however, the whole-body action space -- legs, torso, arms, and hands -- is far higher-dimensional and heterogeneous, raising tokenization, training, and real-time inference challenges that the previous VLA models do not address. We present Holo-M, to our knowledge the first discrete VLA model for humanoid loco-manipulation that intrinsically exploits the language model by extending its vocabulary with action tokens. In this model, we devise a unified action tokenizer that decomposes the humanoid action space into four body-part-specific tokenizers -- end-effector, body, hand, and kinematics -- enabling training across drastically different embodiments and data sources, including humanoid teleoperation, ego-centric human video, and simulation. By extending the language model's vocabulary with these action tokens, we avoid the knowledge-insulation problem inherent to the models that use separate continuous action experts. To meet real-time control requirements, we decode each body part's action tokens through grouped discrete diffusion decoding, rather than using autoregression on the action tokens. We have conducted extensive experiments on the SIMPLE humanoid loco-manipulation benchmark, in which Holo-M achieves the highest success rates in both the generalist and specialist evaluations, leading the second best by significant margins. We will release all the code and model weights.