Get the best builds by email. No account needed.
Live drop
Google DeepMindDropped Oct 27, 2026
Real-time world model adds native audio, live text steering, and coherent multi-step physics with Gemini planning.
About this release
Google DeepMind's latest research world model generates photorealistic interactive 3D environments in real time. It produces native audio (ambient, music, dialogue) alongside visuals, supports natural-language scene steering (camera, weather, characters), and uses Gemini models for coherence and planning. Released as limited research preview with rollout planned for Ultra subscribers and API partners.
Read the official Google DeepMind launch pageFirst builds incoming — this lane fills in automatically as people start building with Genie 4. Know one? Submit the URL.