About
Genie 4 is Google DeepMind's natively multimodal world model, announced October 27, 2026. This independent guide explains what it does, how it differs from Genie 3, and what the official demos actually show: a real-time environment you can steer with plain language, complete with physics, ambient audio and characters that answer you. The site walks through the architecture (Gemini Omni grounding visuals, Gemini 3.8 Flash running the real-time loop), the staged access path from trusted testers to AI Ultra subscribers via Project Genie, SynthID watermarking, and how to write world prompts that steer a scene instead of just describing one. Includes a side-by-side Genie 4 vs Genie 3 comparison, a gallery of source clips from the DeepMind announcement, and an honest list of what Google has not published (frame rate, resolution, model size, pricing).