Google DeepMind Updates Project Genie with Genie 4 and Chainable Worlds
Google DeepMind has released Genie 4, a new interactive 3D world model featuring real-time text editing, native audio, and community remixing tools.

Google DeepMind has announced a major update to Project Genie, introducing Genie 4 alongside new capabilities for its predecessor, Genie 3 12. Genie 4 is designed as the company's most advanced world model for generating interactive 3D environments, allowing users to reshape virtual worlds in real time through text prompts while navigating 12. For the first time, Genie 4 pairs high-fidelity visuals and physics with real-time native audio that adapts dynamically to the unfolding scene, ranging from ambient soundscapes to character dialogue 2. In parallel, Genie 3 has received updates including extended exploration sessions of up to three minutes, 360-degree panorama image grounding, and chainable worlds that allow users to link multiple environments together 12. To streamline creation, Google introduced a Project Genie agent that translates spoken natural language descriptions into customized visual looks and background scores, while integrating Gemini models to offer context-aware suggestions during exploration 2. Additionally, a new community gallery enables users to explore curated experiences—such as the gothic mystery 'Villa Giallo' created with technologist Will (BLVCKL!GHT) and 'The Life of Ants' built with filmmaker Anand Gandhi—and remix existing creations by tweaking underlying structures, soundtracks, or rules 2. Google DeepMind has also implemented safety measures, ensuring all generated video and audio include SynthID watermarking and pass through content safety screenings 2.
Written with AI from the cited sources and checked against them. Open the sources to verify.