Creation & 3DSupported here
Marble
Creates persistent 3D worlds from images, text, video, spatial layouts.
CompanyWorld Labs
OutputPersistent explorable 3D worlds+3 more
AccessPublic productProduct launch
Models and tools
Explore AI World Models
Find a model by what it creates, how visitors can access it, and whether the result is a 3D world, an interactive demo, a physical-AI system, or a research signal.Model directory
Find a model by capability and access.
24 models
Capability
Access
Creation & 3DSupported here
Skybox AI
Generates 360-degree environments: skyboxes, HDRI lighting, VR scenes, game-engine backplates.
CompanyBlockade Labs
Output360-degree equirectangular panoramas+3 more
AccessPublic productPublic app, API documentation, and export workflow
Creation & 3DSupported here
Echo-2
Turns text, images, or panoramas into persistent 3D worlds that can be explored and shared in a browser.
CompanySpAItial
OutputExplorable 3D world+3 more
AccessPublic productPublic web app, gallery, and developer API
Creation & 3DSupported here
SpatialGen
Generates spatially consistent indoor 3D scenes from an image or a text-conditioned layout workflow.
CompanyManycore Tech
OutputNavigable indoor 3D scene+1 more
AccessLocal open sourceOpen model and code with related Aholo API capabilities
Research & adjacent
GWM-1
Runway's general world model spanning generated worlds, avatars, robotics-oriented directions.
CompanyRunway
OutputGeneral world model
AccessPublic productResearch announcement with deployed character-agent surface
CompanyWorld Labs
OutputWorld model platform
AccessPublic productAPI announcement
Interactive worlds
Oasis
Realtime open-world model: interactive video world from keyboard input, frame-by-frame.
CompanyDecart AI + Etched
OutputRealtime open-world AI model
AccessPublic demoPublic demo with open code and model-weight links
Interactive worlds
Odyssey-2
A live generated video world that responds to text while the stream is running.
CompanyOdyssey
OutputInteractive generated video stream
AccessPublic demoPublic interactive experience with ongoing product development
Interactive worlds
Waypoint 1.5
A real-time generated world that can run in a browser or locally on supported consumer hardware.
CompanyOverworld
OutputLive interactive visual world
AccessPublic demoBrowser experience plus open-weight local Biome release
Interactive worlds
WHAM-RT
A real-time gameplay model that generates visuals in response to controller input.
CompanyMicrosoft Research / Xbox
OutputInteractive generated gameplay visuals
AccessPublic demoPlayable Quake II technical demo in Copilot Labs
CompanyAlibaba
OutputExpressive portrait video model
AccessLimited accessResearch project and Alibaba Cloud Model Studio API
Interactive worlds
Genie 3
Frontier world model generating and navigating interactive environments from prompts.
CompanyGoogle DeepMind
OutputInteractive world model
AccessLimited accessResearch preview with Project Genie prototype rollout
Interactive worlds
HappyOyster
Alibaba's open-ended world model for real-time immersive creation, exploration, directing.
CompanyAlibaba
OutputRealtime explorable worlds+2 more
AccessLimited accessEarly Access program
Creation & 3D
MoWorld 3D
A world-model platform for dynamic scene reconstruction, editing, novel views, and physical-AI simulation.
CompanyKOKONI / MoWorld
OutputInteractive 3D scene+1 more
AccessLimited accessOfficial product launch with limited public self-service access
CompanyNVIDIA
OutputWorld foundation model platform
AccessLocal open sourcePlatform launch: Cosmos 3 preview, data-factory blueprint, active GitHub stack
Physical AI
HY-Embodied-0.5
Tencent's embodied foundation-model family: spatial reasoning, planning, robot perception, 0.5-X.
CompanyTencent Hunyuan
OutputEmbodied foundation model
AccessLocal open sourceOpen-source model family with enhanced 0.5-X release
Creation & 3D
HY-World 2.0
Reconstructs, generates, simulates persistent 3D worlds from text, images, video.
CompanyTencent Hunyuan
OutputGenerated 3D worlds+4 more
AccessLocal open sourceOpen-source report, weights, and May 2026 world-generation code expansion
Physical AI
LingBot-Map
Feed-forward 3D model for streaming video reconstruction; public KITTI/Oxford-Spires benchmarks.
CompanyAnt Group / Robbyant
OutputStreaming 3D foundation model
AccessLocal open sourceOpen-source code, paper, model releases, and public evaluation scripts
CompanyAnt Group / Robbyant
OutputRobot-control world model
AccessLocal open sourceOpen-source code, paper, and model releases
Physical AI
LingBot-VLA
Robbyant's pragmatic vision-language-action model for generalist robotic manipulation across platforms.
CompanyAnt Group / Robbyant
OutputVision-language-action foundation model
AccessLocal open sourceOpen-source code, paper, and model releases
Interactive worlds
LingBot-World
Robbyant's open-source world simulator for high-fidelity, long-horizon, real-time interactive environments.
CompanyAnt Group / Robbyant
OutputOpen-source interactive world simulator
AccessLocal open sourceOpen-source code, paper, and model releases
Interactive worlds
Matrix-Game 3.0
A streaming keyboard-and-mouse world model designed for real-time interaction and long-horizon memory.
CompanySkywork AI
OutputStreaming generated video world
AccessLocal open sourceOpen-source code and model release
Research & adjacent
Project Sid
Minecraft-based many-agent simulation: specialization, law, culture, religion in AI civilizations.
OutputMany-agent civilization simulation
AccessNo consumer accessPublic research report and GitHub project
Physical AI
V-JEPA 2
A video-trained latent world model for understanding, predicting, and planning actions in the physical world.
CompanyMeta
OutputLatent video representations+2 more
AccessNo consumer accessOpen code, checkpoints, paper, and physical-reasoning benchmarks