Trusence Every claim has a source
Last updated 21 November 2026 Search Türkçe
Models

Google opens Genie 4 world model to Ultra subscribers and developers

Users can now change a running world with text prompts, while developers get the same model through the Gemini API and Google AI Studio.

Updated 19 November 2026

Google has opened its Genie 4 world model to eligible Google AI Ultra subscribers through a revamped Project Genie experience, and to developers via the Gemini API and Google AI Studio. Genie 4 lets users alter a running world with new text prompts, adding weather changes, characters or events, and it produces synchronized native audio. Genie 3 also gained longer sessions of up to three minutes per generation, world chaining, and seeding from 360-degree panoramas, while a new agent and community gallery help users build and remix worlds. On the developer side, Genie 4 generates continuous 720p worlds at 24fps from a text prompt, a reference image or both, with pricing details published separately. Google says all outputs carry SynthID watermarking and that its safety screenings cover model inputs and outputs.

Why it matters

For people following AI, the release moves interactive world generation from research demos into products that ordinary subscribers and developers can use. Users can steer a generated world as it runs rather than only at the start, and every output carries SynthID watermarking, which gives a built-in way to trace AI-generated media. Taken together, the changes suggest that generated environments with synchronized audio are becoming a standard offering rather than an experiment.

Sources · 2