egov.mn
Монголоор унших

Google DeepMind opens Genie 4 world model to users with real-time world editing and native audio

Google is rolling out a new Project Genie experience to eligible Google AI Ultra subscribers, giving first access to Genie 4, its most realistic interactive world model, along with a creation agent and a community gallery.

This story was prepared automatically by software.

Google Project Genie promotional image reading “Step inside your imagination”
Google Project Genie promotional image reading “Step inside your imagination”

Google DeepMind and Google Labs on November 18 rolled out a new version of Project Genie, their experimental prototype for generating and exploring interactive worlds, giving users first-time access to Genie 4. The update is available to eligible Google AI Ultra subscribers, while Google AI Pro members can join a waitlist.

Genie 4: real-time edits and native audio

Google describes Genie 4 as its most realistic and interactive world model yet. Users can now type a new prompt mid-session to change their world in real time, such as shifting the weather, dropping in a character or redirecting the action. For the first time, the visuals are paired with real-time native audio that adapts as the scene unfolds, from ambient soundscapes to character dialogue.

Genie 3 remains available and has also been upgraded: sessions now last up to 3 minutes per generation, up from 60 seconds, worlds can be chained together, new worlds can be seeded from 360-degree panorama images, and users can trigger in-world events with natural-language prompts.

A creation agent and a community gallery

A new Project Genie agent turns a natural-language description into a world, orchestrating Google's multimodal models to create the initial visuals and a matching background score. Users can also bring Gemini into a session to observe what is on screen, suggest where to go next or trigger in-world events. A new gallery lets people publish their worlds and explore and remix others' creations.

Responsibility measures

Google said all video and audio generated in Project Genie carries SynthID watermarking and that all model inputs and outputs go through safety screenings. The company stressed that Genie 3 and Genie 4 remain frontier research models and that it is still working on challenges including camera latency, intricate physical interactions, real-time audio precision and long-term scene permanence.

Sources: Google

Share
Next storyOregon sues OpenAI for at least $2 billion, alleging ChatGPT endangers children
Newsletter

Get the latest news delivered to your inbox.

Related articles