Google announced its Genie 4 world model, designed to generate interactive experiences using artificial intelligence. According to the company, the updated model introduces several key improvements, including more realistic worlds, real-time updates through natural language prompts, audio generation, and the integration of Gemini 3.8 Flash to support simulations.
Upgraded Features and Capabilities
The new model extends the duration of generated experiences to roughly two minutes, doubling the one-minute limit of the previous Genie 3 version. Google positions the software as a practical utility for filmmaker pre-visualization, game prototyping, and the creation of robotic training scenarios. However, the company cautions that as simulations continue to run longer, fine details and object permanence may change over time.
Google is initially rolling out Genie 4 as a research preview directed at trusted testers. The company stated that the model will arrive soon for Google AI Ultra subscribers, and a waitlist is currently open for users wishing to test the technology through Project Genie.
