Motus: A Unified Latent Action World Model

TL;DR

Motus is a unified latent action world model that integrates understanding, generation, and control using a Mixture-of-Transformer architecture, enabling flexible multimodal modeling and large-scale pretraining for robotic tasks.

Contribution

It introduces a novel Mixture-of-Transformer architecture and a comprehensive training pipeline for unified multimodal world modeling in embodied agents.

Findings

Achieves +15% to +45% performance improvements over state-of-the-art methods.

Demonstrates effective large-scale pretraining of latent actions.

Improves downstream robotic task performance in simulation and real-world scenarios.

Abstract

While a general embodied agent must function as a unified system, current methods are built on isolated models for understanding, world modeling, and control. This fragmentation prevents unifying multimodal generative capabilities and hinders learning from large-scale, heterogeneous data. In this paper, we propose Motus, a unified latent action world model that leverages existing general pretrained models and rich, sharable motion information. Motus introduces a Mixture-of-Transformer (MoT) architecture to integrate three experts (i.e., understanding, video generation, and action) and adopts a UniDiffuser-style scheduler to enable flexible switching between different modeling modes (i.e., world models, vision-language-action models, inverse dynamics models, video generation models, and video-action joint prediction models). Motus further leverages the optical flow to learn latent actions and adopts a recipe with three-phase training pipeline and six-layer data pyramid, thereby extracting pixel-level "delta action" and enabling large-scale action pretraining. Experiments show that Motus achieves superior performance against state-of-the-art methods in both simulation (a +15% improvement over X-VLA and a +45% improvement over Pi0.5) and real-world scenarios(improved by +11~48%), demonstrating unified modeling of all functionalities and priors significantly benefits downstream robotic tasks.