AI tools for World Models

ScribVet is an AI-powered veterinary scribe tool designed to streamline the documentation process for veterinarians, allowing them to focus more on patient care rather than admi

AI Song Maker is an advanced music creation platform powered by artificial intelligence, designed to enable users to generate high-quality songs effortlessly. The platform provi

AltTextLab is an advanced solution designed to automate the creation of descriptive and compliant alternative text for images across the web. By leveraging a sophisticated visio

Genie 4 is Google DeepMind's interactive world-simulation model. It generates explorable environments with coordinated visual motion and native audio, aiming to let users move t

WorldSculpt reconstructs complex scenes as collections of complete object meshes that can be edited separately. It targets densely cluttered environments containing hundreds of

Code World Model uses a coding agent to update and control the state of a generated world. It separates persistent world structure from visual appearance, allowing the environme

H3-World turns keyboard-style character and camera actions into controllable video dynamics. It builds on MiniMax-H3 to generate visually coherent environments that follow chang

OpenWAM is a modular research stack for studying how video world models can learn executable robot control. It separates infrastructure, controlled design experiments, and the r

SolarWM is an open framework for training video world models that support long-horizon, camera-controlled interaction. It combines a large video-data pipeline with training and

OVOW, or One Video One World, reconstructs a monocular video as a structured 4D scene with separate mesh objects and motion. It targets simulation-ready output rather than only

UnifoLM-WLA-1.0 is Unitree Robotics general-purpose humanoid foundation model for coordinated manipulation. A single six-billion-parameter model handles desktop and whole-body t

Masked Visual Actions is a unified world modeling project for robot video control using masked visual action representations. It frames action as a visual signal that can suppor

Wonder is an interactive video world model for image- and video-conditioned world generation. It builds persistent environments that users can navigate with camera movement whil

PhiZero is a research world model built around physical language, a compact discrete representation of world-state transitions learned from in-the-wild video. It separates physi

ShadowDancer is a video world model research project that teaches a single model many different actions by learning unified dynamics representations from synchronized shadow pai

Open-Dreamer is an open project from Next State about training a frontier-level world model. It provides a practical account of model design, training lessons, failures, fixes,

Modality Forcing is a post-training recipe that turns a pretrained text-to-image model into a joint image-depth generator. It leverages the spatial priors inside T2I models to s

LabOS is an AI-XR co-scientist platform for modern laboratories that connects agentic reasoning with real-world experiment execution. It is designed to help scientists move from

MolmoAct 2 is an open robotics foundation model from Ai2 designed to help robots perform real-world manipulation tasks with stronger perception and action reasoning. It focuses

Multiworld is a scalable multi-agent, multi-view video world model designed to simulate environments where several agents act at the same time and must remain consistent across

Lyra 2 is an NVIDIA research system for explorable generative 3D worlds. It is designed to create camera-controlled walkthrough videos and lift them into 3D through feed-forward

Waypoint 1.5 is a real-time AI world model from Overworld designed to bring interactive generated worlds to everyday GPUs. It focuses on making world simulation more accessible

HY World 2.0 is Tencent Hunyuan's open-source multimodal world model for reconstructing, generating, and simulating 3D worlds. It is designed as a broad world-modeling framework