To Their Credit

Google opens Genie 4 interactive worlds to developers

The Gemini API release brings live, editable 720p simulated environments with native audio into Google AI Studio.

Ledger: AdoptedAI is taken up

Google has opened its Genie 4 world model to developers through the Gemini API and Google AI Studio, a day after making it available to eligible Google AI Ultra subscribers through its consumer-facing Project Genie tool.

As we reported on Nov. 18, Google had released Genie 4 in Project Genie for eligible AI Ultra subscribers; the new step gives builders access through the API’s Interactions API and through Google AI Studio.

The model generates continuous interactive worlds at 720p and 24 frames per second from a text prompt, a reference image, or both. Developers can move through those environments using keyboard controls, then alter weather, lighting, objects and events through live text prompts during a session.

Genie 4 produces synchronized native audio alongside its visuals, including ambient soundscapes, sound effects and character dialogue. Google says Gemini reasoning is intended to maintain the generated world’s state and coherence as a session unfolds.

The company is positioning the release for prototypes of interactive experiences, spatial-design visualisation, and robotics or AI-agent training. Google AI Studio offers an interactive canvas and starter apps, while the company has directed developers to Gemini API documentation and pricing information.

Every Genie 4 output carries an imperceptible SynthID watermark. SynthID is Google’s system for identifying AI-generated or AI-altered material, using invisible marks in images and video and inaudible marks in supported audio.

The availability does not settle technical limits associated with simulated worlds. Superpower Daily noted that latency, intricate physical interactions, precise real-time audio and maintaining scenes over long periods remain unresolved challenges. Google has similarly characterised Genie 3 and Genie 4 as frontier-research world models while saying it is working on those constraints.

The story so far

  1. Google opens Genie 4 interactive worlds to AI Ultra subscribers
  2. Google opens Genie 4 interactive worlds to developers This post

See also

More news

All news
  1. Vietnam journalism seminar puts human judgment at centre of AI trainingLedger: Disputed
  2. Mikami and Kim split over AI’s role in game developmentLedger: Disputed
  3. Tamil Nadu arrests publisher in alleged AI library-book procurement fraudLedger: Disputed