Google Project Genie

Full Description

Project Genie is Google DeepMind's experimental AI-powered world model that generates interactive 3D environments from text or image prompts. Powered by Genie 3, it creates real-time, explorable worlds at 720p/24fps with consistent physics and memory lasting several minutes. Key features include text-to-world generation using natural language prompts, image-to-world conversion from photos, artwork, or sketches, real-time generation as you navigate (not pre-rendered), and integration with Nano Banana Pro for world sketching. Use cases include rapid game prototyping and level design, AI/robotics training environments, educational simulations, and architecture visualization. Google's SIMA agent already uses Genie 3 for task learning. Currently available through Google Labs for Google AI Ultra subscribers ($249.99/month) in the US. While still a research prototype with 60-second session limits, it represents a significant advance in AI-generated interactive environments.