world-models.io — The Knowledge Hub for AI World Models
world-models.io is a structured knowledge platform covering AI world models — from model-based reinforcement learning to foundation world models, embodied AI, and simulation engines.
Explore world models like DreamerV3, Genie 3, V-JEPA 2, NVIDIA Cosmos, MuZero, Sora, DIAMOND, OASIS, LeWorldModel, PixVerse R1, and more. Compare architectures, read research, and discover the labs building the future of AI.
Featured World Models
Genie 3 — Google DeepMind's real-time interactive 3D world generation from text at 24fpsV-JEPA 2 — Meta FAIR's self-supervised world model for visual understanding and zero-shot robot controlDreamerV3 — General algorithm mastering diverse domains through world model learningNVIDIA Cosmos — World foundation model platform for physical AIGenie 2 — Interactive 3D environment generation from single imagesSora — OpenAI's video generation model with emergent world understandingMuZero — Superhuman game play through learned planningLeWorldModel — Compact 15M-parameter JEPA that learns physics on a single GPUV-JEPA — Self-supervised visual world model by Meta FAIRDIAMOND — Diffusion-based world model for reinforcement learningOASIS — Open-source real-time neural game enginePixVerse R1 — First multiplayer real-time world model with shared worldsTD-MPC2 — Scalable world model agent for 104+ tasksIRIS — Autoregressive world model bridging language modeling and RLGAIA-1 — Generative world model for autonomous drivingCopilot4D — 4D LiDAR world model for self-drivingI-JEPA — Image-based Joint Embedding Predictive ArchitectureGameNGen — Neural game engine simulating DOOM in real-time3D-VLA — 3D vision-language-action model for embodied AIPandora — Generative world model for open-ended environmentsPlaNet — Deep Planning Network for latent space controlRSSM — Foundational architecture behind the Dreamer familyDreamerV2 — First model-based agent to achieve human-level AtariWorld Models (Ha & Schmidhuber) — The seminal 2018 paper that popularized world modelsPredictron — End-to-end learning and planning via abstract world modelsUniSim — Universal simulator for real-world interactionsI2A — Imagination-augmented agents for deep RLVPN — Value Prediction Network for abstract planningEmu Video — Factorized video generation via image conditioningAMI World Model — Multimodal foundation model for embodied AIRT-2 — Vision-Language-Action model for robotic controlGen-3 Alpha — Advanced video generation by RunwayStable Video Diffusion — Open-source video generation foundation modelMILE — Model-based imitation learning for autonomous drivingLWM — Large World Model processing 1M+ tokens of video and textGenie — First generative interactive environment from unlabeled videoSTEVE-1 — Instruction-following agent for Minecraft
Explore
Research Topics
Comparisons
Guides
Labs & Organizations
Categories
Knowledge Hub