Google Genie 4: AI World Model Gets Sound in 2026

By PowerfulAI.tools Editorial Team

Google DeepMind has announced Genie 4, the newest version of its "world model", an AI system that generates an interactive 3D-style world you can move through in real time. The headline change is sound: Google says Genie 4 is its first real-time world model to generate audio natively, so every world comes with ambient sound, music and spoken dialogue made on the fly.

Here is what was announced on 27 October 2026, who can use it, and what it means if you are not a developer or researcher.

What launched

Genie 4 creates a live, explorable scene from a prompt and keeps generating it as you move. Google describes it as an "early, experimental model" that builds on its Gemini family. According to Google DeepMind, Gemini Omni helps keep the visuals grounded in how things look and behave in the real world, while other Gemini models handle reasoning so the scene stays coherent.

The main features Google lists are:

• Native audio: ambient sound, music and interactive dialogue are generated alongside the visuals.
• Steering with plain language: you can change a running scene with a text prompt, such as changing the weather, moving the camera or adding a person or object.
• More realistic motion: Google points to simulated aerodynamics and collisions, and characters that move naturally and can be steered in real time.
• Keyboard control: testers used standard keyboard controls to move around and shift the camera, plus text prompts.
• Watermarking: Google says all generated video and audio will include SynthID, its invisible watermark for identifying AI-generated content.

Google also says Genie 4 is meant for more than entertainment. It names uses such as safe training environments for robots and AI agents, and creative tools for filmmaking, journalism, gaming and manufacturing.

Who can use it, and where

For now, access is very limited. Google says Genie 4 "has been released as a research preview to trusted testers". It is described as "coming soon to Google AI Ultra subscribers in Project Genie", along with early access partners in the Gemini API and Google AI Studio. Google has not given a date for the wider rollout. There is a waitlist form linked from the Google DeepMind announcement.

Project Genie is Google's existing experimental app for creating and exploring AI-generated worlds. When it launched in January 2026, Google said it was rolling out to Google AI Ultra subscribers in the US aged 18 and over, with more territories "in due course". The current Project Genie page says it is available to Google AI Ultra subscribers aged 18+. Google has not said which countries will get Genie 4 when it reaches Project Genie, so Australian users should not assume access yet.

What it costs

Google has not announced a separate price for Genie 4. The only plan it names is Google AI Ultra, its top-tier AI subscription. If you are weighing up Ultra for Genie 4, check the current price and availability on Google's official Google AI plans page for your country, and keep in mind that the Genie 4 rollout to Ultra has no confirmed date.

The limits Google admits

Google is upfront that this is an early model. Its announcement says visual details and object permanence can drift, especially in longer sessions, after several prompt edits, or while moving through a scene. Complex physical interactions, real-time audio accuracy, legible text, spatial accuracy and long-term consistency are all listed as ongoing research areas.

What it means for ordinary users

For most people, Genie 4 is something to watch rather than something to use this week. It is not a general chatbot and it is not available on free or mid-tier plans. Its importance is in where it points: AI that does not just make a fixed video clip, but a world you can walk around in, talk to and change as you go, with sound.

If you make videos, games or creative prototypes, it is worth joining the waitlist and following Project Genie. If you simply want to try AI world-building today, Project Genie's earlier version is the official route, but only on Google AI Ultra and only where Google offers it.

How we checked

We confirmed every detail in this post on 28 October 2026 using Google's own pages only:

• Google's announcement: https://blog.google/innovation-and-ai/models-and-research/google-deepmind/introducing-genie-4/
• Google DeepMind's Genie 4 post: https://deepmind.google/blog/genie-4-a-new-generation-of-world-simulation/
• Project Genie launch post: https://blog.google/innovation-and-ai/models-and-research/google-deepmind/project-genie/
• Project Genie page: https://labs.google/projectgenie

We have not tested Genie 4. Pricing, countries and release dates that Google has not published are left out.