New: the Timeline is live. Track world model releases, papers, and benchmark updates in real time.
world-models.io
The Knowledge Hub for AI World Models

World Models Relationship Graph

A relationship map connecting AI world models to their research labs and technical categories, with a static entity directory available before JavaScript runs.

robotics model-based-rl simulation embodied-ai

Quick Answer

Short extractable summary preserved directly in static HTML.

  • The knowledge graph connects 48 model records with 32 labs and 8 technical categories.
  • The interactive canvas is an enhancement; the complete linked entity directories below are delivered in the initial HTML for crawlers, assistive technology, and no-JavaScript readers.

Model Nodes

ModelLab or companyCategory
DreamerV3Google DeepMindModel-Based RL
DreamerV2GoogleModel-Based RL
PlaNetGoogleModel-Based RL
RSSMGoogleLatent Dynamics
NVIDIA CosmosNVIDIAFoundation World Model
Genie 2Google DeepMindGenerative World Model
MuZeroGoogle DeepMindModel-Based RL
PredictronGoogle DeepMindModel-Based RL
UniSimGoogle DeepMindGenerative World Model
TD-MPC2MIT / MetaModel-Based RL
World Models (Ha & Schmidhuber)Google Brain / IDSIAModel-Based RL
V-JEPAMetaSelf-Supervised World Model
IRISMicrosoft ResearchModel-Based RL
GAIA-1WayveFoundation World Model
Imagination-Augmented Agents (I2A)Google DeepMindModel-Based RL
Value Prediction Network (VPN)University of Michigan / Google BrainModel-Based RL
AMI World ModelAMI LabsFoundation World Model
SoraOpenAIGenerative World Model
DIAMONDMicrosoft Research / University of GenevaModel-Based RL
PandoraTsinghua University / ByteDanceGenerative World Model
OASISDecart / EtchedGenerative World Model
Copilot4DWaabiFoundation World Model
GenieGoogle DeepMindGenerative World Model
I-JEPAMetaSelf-Supervised World Model
GameNGenGoogle ResearchGenerative World Model
Emu VideoMetaGenerative World Model
3D-VLAMIT / TsinghuaFoundation World Model
RT-2Google DeepMindFoundation World Model
Large World Model (LWM)UC BerkeleyFoundation World Model
Stable Video DiffusionStability AIGenerative World Model
MILEWayveFoundation World Model
STEVE-1UT AustinGenerative World Model
Gen-3 AlphaRunwayGenerative World Model
Genie 3Google DeepMindGenerative World Model
V-JEPA 2MetaSelf-Supervised World Model
LeWorldModelMila / NYU / Samsung SAILSelf-Supervised World Model
PixVerse R1PixVerseGenerative World Model
MarbleWorld LabsFoundation World Model
1X World Model1XFoundation World Model
PlayWorldPrinceton UniversityGenerative World Model
WHAMMicrosoft Research / Ninja TheoryGenerative World Model
WHAM-RTMicrosoft Research / Ninja TheoryGenerative World Model
GAIA-2WayveGenerative World Model
Waabi WorldWaabiGenerative World Model
Odyssey-2OdysseyFoundation World Model
HY-World 2.0Tencent HunyuanFoundation World Model
RELICAdobe ResearchGenerative World Model
Matrix-Game 2.0Skywork AIGenerative World Model

Lab Nodes

LabTypeFocus
Google DeepMindindustrymodel-based-rl, generative-world-models, planning, games, robotics
NVIDIA Researchindustryfoundation-world-models, physical-ai, autonomous-driving, robotics, video-generation
Meta FAIRindustryself-supervised-learning, jepa, embodied-ai, autonomous-intelligence
Mila / NYU / Samsung SAILacademicself-supervised-learning, jepa, physical-reasoning, embodied-ai
AMI Labsindustrymultimodal-world-models, embodied-ai, language-conditioned-robotics, foundation-models
UC Berkeleyacademicmodel-based-rl, robotics, visual-rl, sim-to-real
Stanford Universityacademicrobotics, embodied-ai, manipulation, simulation
Carnegie Mellon Universityacademicrobotics, autonomous-systems, model-based-learning
Toyota Research Instituteindustryautonomous-driving, robotics, physical-ai
MIT CSAILacademicmodel-based-rl, planning, embodied-ai, scalable-agents
OpenAIindustryvideo-generation, world-simulation, foundation-models, generative-ai
Wayveindustryautonomous-driving, generative-world-models, end-to-end-driving, scenario-generation
PixVerse Researchindustryreal-time-world-models, interactive-generation, multiplayer-experiences, video-generation
Decartindustryreal-time-world-models, interactive-generation, neural-game-engines
Waabiindustryautonomous-driving, lidar-world-models, 4d-prediction, closed-loop-simulation
Microsoft Researchindustryautoregressive-world-models, diffusion-rl, sample-efficient-rl
Google Researchindustryneural-game-engines, world-simulation, video-understanding
Stability AIindustryvideo-generation, diffusion-models, open-source-ai
Runwayindustryvideo-generation, creative-ai, controllable-generation
UT Austinacademicinstruction-following, open-world-agents, video-pretraining
World Labsindustry3d-generation, multimodal-world-models, persistent-worlds, spatial-reasoning
1Xindustryhumanoid-robotics, physical-ai, video-world-models, embodied-ai
Princeton Universityacademicrobotics, world-models, manipulation, autonomous-data-collection
Tsinghua Universityacademicembodied-ai, generative-world-models, robotics, 3d-reasoning
ByteDance Researchindustrygenerative-world-models, video-generation, embodied-ai
University of Genevaacademicmodel-based-rl, diffusion-models, world-models
IDSIAacademicfoundational-world-models, model-based-rl, generative-world-models
Etchedindustryreal-time-world-models, interactive-generation, neural-game-engines
Odysseyindustryinteractive-world-models, causal-video, real-time-generation
Tencent Hunyuanindustry3d-world-models, multimodal-generation, world-reconstruction
Adobe Researchindustryinteractive-world-models, long-horizon-memory, video-generation
Skywork AIindustryinteractive-world-models, real-time-generation, open-source

Category Nodes

CategoryDefinition
Model-Based RLModel-based reinforcement learning is an approach where agents learn a predictive model of the environment (a world model) and use it to simulate outcomes, plan ahead, and learn from imagined experience.
Embodied AIEmbodied AI concerns AI systems that interact with and learn from the physical world through a body, whether a robot, an autonomous vehicle, or a virtual agent with physical constraints.
Generative World ModelsGenerative world models are AI systems that learn to generate realistic simulated environments, replacing or augmenting hand-crafted simulators with learned models of world dynamics.
Latent DynamicsLatent dynamics models learn compressed representations of environment dynamics in a latent (hidden) space, enabling efficient prediction and planning without operating in pixel space.
Autonomous AgentsAutonomous agents use world models to make independent decisions in complex, open-ended environments by predicting consequences and planning actions.
Video World ModelsVideo world models understand and generate video as a representation of world dynamics and physics, learning temporal structure, object permanence, and physical interactions from video data.
Foundation World ModelsFoundation world models are large-scale, general-purpose models trained on massive datasets to learn broad representations of world dynamics. They aim to serve as versatile base models for diverse downstream tasks, from robotics to autonomous driving to video generation.
Self-Supervised World ModelsSelf-supervised world models learn representations of environment dynamics without explicit labels or reward signals. They leverage prediction in abstract representation space, predicting future states, video frames, or embeddings, to build internal models of how the world works.