World models

Research

HAIC: Humanoid Agile Object Interaction Control via Dynamics-Aware World Model

Research

Official Implementation of "Kinema4D: Kinematic4D World Modeling for Spatiotemporal Embodied Simulation"

Research

LiLa-WAM: Lightweight Latent Reasoning World-Action Model for Robotic Manipulation

Research

Source codes for the paper "MindJourney: Test-Time Scaling with World Models for Spatial Reasoning"

Research

MotuBrain: An Advanced World Action Model for Robot Control

Research

Official implementation of NavMorph: A Self-Evolving World Model for Vision-and-Language Navigation in Continuous Environments (ICCV'25).

Research

OpenDW provides an action-conditioned world-model project for embodied intelligence.

Research

Official code for "OpenWAM: An Open Framework for Composable World-Action Models" (Stanford University)

Research

PointWorld studies 3D world models for robotic manipulation.

Research

RynnWorld-4D: 4D Embodied World Models for Robotic Manipulation

Research

repository for training action-conditioned latent diffusion world models for robot video generation

Research

Offical implementation of ICLR'26 paper "Empowering Multi-Robot Cooperation via Sequential World Models"

Research

UnifoLM-WMA-0: A World-Model-Action (WMA) Framework under UnifoLM Family

Research

WAVE-Go: Go2-W world-model action adaptation, verified mapless charging, and Cosmos3-Edge hybrid long-range navigation demos

Research

WMNav: Integrating Vision-Language Models into World Models for Object Goal Navigation

Research

WoW (World-Omniscient World Model) is a generative world model trained on 2 million robotic interaction trajectories, designed to imagine, reason, and act in the physical world. Unlike passive…