Google DeepMind Unveils Genie 4, a World Model That Creates Sound in Real Time

Google DeepMind Unveils Genie 4, a World Model That Creates Sound in Real Time

Google DeepMind has unveiled Genie 4, a new AI world model that builds interactive scenes in real time and generates sound natively as it goes. DeepMind calls it the first real-time world model with native audio: every world gets its own ambient sound, such as wind or waves, and characters inside the scene can talk and react to the user or to events around them.

Users can reshape a scene with plain text, changing the weather, the camera angle, or adding new people and objects while the world keeps running. Gemini 3.8 Flash works as a guiding agent during generation, splitting a goal into steps, checking each step, and changing the plan when the user steps in. DeepMind has not published figures for resolution, frame rate, or session length, and the model is currently a research preview for trusted testers rather than a public release.

Google AI Ultra subscribers will get Genie 4 in Project Genie, while early access partners will receive it through the Gemini API and Google AI Studio, with no date given. The launch came on the same day Oracle expanded its cloud to offer Google TPUs, Cerebras chips, and multi-gigawatt Nvidia Vera Rubin clusters side by side, and Apple added AI assistants to Pages, Keynote, and Numbers. Together, the announcements show AI moving deeper into the tools people already use, rather than arriving only as new standalone apps.

Source: read more

Comments

Popular posts from this blog

Fairbairn's Walk-Off Field Goal Seals Wembley Thriller as Texans Edge Jaguars 34-31

Anthropic IPO: Public S-1 Filing Expected Today Ahead of $2 Trillion Listing

Netflix Beats Earnings but Slashes Ad Forecast — Stock Tumbles After Hours