What changed: According to Ars Technica and The Decoder, Google DeepMind introduced Genie 4, the successor to Genie 3. It generates interactive worlds frame by frame in real time from text prompts and is DeepMind's first world model with native audio, including ambient sound and character dialogue. Gemini models, including Gemini 3.8 Flash, plan and steer the simulation in the background to keep it coherent, and the company says physics and human movement look more realistic.
Who is affected
Per the sources, only select trusted testers can access it for now; it will later expand to developers via the Gemini API and Google AI Studio, and to Google AI Ultra subscribers.
What it means for your business
For businesses using the Gemini API or Google AI Studio, this means they may eventually be able to build real-time, audio-enabled simulation environments through Genie 4; pricing has not been announced yet.
What to watch
No action is required right now. Once Genie 4's wider access and pricing are announced, businesses using the Gemini API or AI Studio should evaluate whether the tool fits their workflow.
Timing
The sources give no firm date. Genie 4 is currently in use with select testers; no timeline has been announced for wider access for Google AI Ultra subscribers or Gemini API/AI Studio partners.
Source (press reports): Ars Technica, The Decoder
