Google launches Genie 4, a world model powered by Gemini 3.8 Flash designed to generate interactive AI simulations, initially as a preview to trusted testers

You can generate longer, more complex interactive AI simulations for creative and training purposes.

Google has launched Genie 4, an AI-powered world model capable of generating interactive simulations. This new iteration features more realistic environments, real-time updates via natural language prompts, integrated audio generation, and is powered by Gemini 3.8 Flash. Genie 4 can create simulations up to two minutes long, an increase from its predecessor's one-minute limit. Potential applications include pre-visualization for filmmakers, game idea prototyping, and robotic training scenarios. Google cautions that prolonged simulations may exhibit shifts in detail and object permanence. The model is initially available as a research preview for trusted testers and will soon be accessible to Google AI Ultra subscribers. A waitlist is open for those wishing to try it via Project Genie.

Google's Genie 4 signifies a leap in generative AI for interactive simulations, moving beyond static content to dynamic, responsive environments. The integration of Gemini 3.8 Flash and enhanced realism suggests a trend towards more sophisticated AI models that can understand and respond to complex, real-time user input, potentially blurring the lines between creation tools and interactive agents.

This development could significantly impact fields like game development, film pre-visualization, and robotics training by lowering the barrier to creating detailed simulations. As Google offers this as a preview, it signals an industry-wide push for more accessible and powerful generative tools, hinting at future AI capabilities that can produce richer, more interactive digital experiences for a wider range of users.