Genie 4 Explained: Google's World Model and Who Can Use It

Who can try Genie 4 in Project Genie, how live editing and audio work, and when creators and developers should use it for a prototype.

Published

Genie 4 Explained: Google's World Model and Who Can Use It

Genie 4 is Google DeepMind’s world model. You can step into a scene, turn a corner and ask for rain while you’re still walking. In Project Genie, eligible Google AI Ultra subscribers can try it now, and Google AI Pro members can join the waitlist.

What happens when you steer Genie 4?

A world model predicts how a place will change in response to what you do. Turn left at a fork and it generates the view down the left passage, taking its cue from where you were standing and which way you turned.

Google’s The Life of Ants, made with filmmaker Anand Gandhi, puts you at the height of a worker ant. You can follow a trail, collect honeydew from aphids and carry food back to the colony. In a nature film the filmmaker picks the route. Here the interesting question is what you would stop to look at along the way.

A clay ant at a fork illustrates how each movement changes the next view.
Each view gives you another choice about where to go.

You can also change the scene from inside it. Type a request to change the weather or bring in a character while you explore. Genie 4 generates the sound as the scene unfolds, including environmental sounds and character dialogue. That’s what Google means by native audio. The music that the Project Genie creative agent can compose when it helps you set up a world serves a different purpose. The agent helps establish the mood. The model responds to what happens on the walk. Google’s November 18 launch announcement

Who can use Project Genie?

Eligible Google AI Ultra subscribers can enter now. Google AI Pro members can join a waitlist. Google’s access instructions, checked November 19, 2026, specify a personal Google Account, age 18 or older, in a country where Ultra is available. Google emails Pro members as places open. The announcement names no access route for Plus or free accounts. Anyone curious about the launch film can start with the short previews on the public gallery page.

In the US, Google AI Ultra starts at $99.99 per month. Google’s plan prices · Our guide to Gemini pricing

Project Genie’s public page with its Explore now entry and examples from the gallery.
Project Genie

How to start a world

  1. Open the public gallery

    Preview The Life of Ants to see what exploring from a character’s perspective looks like. To take the controls, choose Explore now and sign in with your personal Google Account. If you have Pro and see Join waitlist, that is your next step.

  2. Give the scene somewhere to go

    With access, open a gallery experience or choose Create with Genie. Describe a place, optionally add a reference image, select Genie 4 and choose Create. For a first prompt, try “A covered flower market just before opening, with a narrow path between the stalls.” The path gives you somewhere to walk and the stalls give you landmarks.

  3. Change one thing while moving

    Use WASD to move and the arrow keys to look around, unless the experience gives you its own controls. Try “Rain starts on the canvas roof.” Listen as you walk from under the awning into the open. For a creative review, ask which part of the change someone noticed first. You can download a video of that playthrough to discuss the same moment later.

When to choose Genie 3

Pick Genie 4 when the scene depends on sound, such as footsteps that announce a character before they appear. Google describes it as the stronger model for visual and physical detail. Both models now accept prompts while you move.

Pick Genie 3 in the same model picker when the starting point is a 360-degree panorama or the route needs a longer run. Google raised its limit from 60 seconds to 3 minutes per generation and added the ability to link worlds. Those are Genie 3 updates. A designer sketching a tour from a panoramic location photo has a reason to choose the older model.

What can a business build with it?

Creators can use it for shot planning. Give an art director a route through a proposed set and ask where they would put the camera. Genie 4 differs from a video generator because the person exploring still controls the viewpoint. A downloaded playthrough preserves one route as a fixed clip.

Game designers can try an encounter before building the level. A world model serves a different job from a game engine. If a puzzle requires a key to stay behind a particular door after the player leaves, an engine can store that fact as a rule. Genie predicts what should appear on returning. DeepMind reports that details can drift as you navigate or edit, and Google documents a delay between input and movement. Use the walk to decide where the key belongs, then implement the rule in your engine.

An illustrated return visit to a clay tunnel shows the same doorway but a different root beside it.
A continuity check for a prototype: leave a landmark, return and check what stayed put.

Product teams can build a prototype through the Gemini API. On November 19 Google opened Genie 4 to developers through the Interactions API and Google AI Studio. This gives a team a way to try a guided showroom or an interactive set from its own software. Before promising customers a faithful tour, walk past a landmark, change the scene, then come back. If the exact layout, product dimensions or repeatable behavior matter to the sale, keep those in a system you control. Google’s developer access announcement

For an app that responds to a live visual feed, compare real-time video generators for interactive apps. If the deliverable is an edited commercial, compare AI video generators for finished clips.

Common questions

When was the Genie 4 release date?

DeepMind announced the model on October 27, 2026. The browser rollout began November 18 for eligible Ultra subscribers, followed by developer access on November 19.

What is Genie 4?

It’s the Google DeepMind model that generates an interactive scene as you explore. Project Genie is the browser experience where you can use it. If you’re looking for Google’s model for coding and agent work, that’s a separate product covered in the Gemini 4 Hydrogen explainer.

Is Genie 3 available to the public?

People with Project Genie access can select Genie 3 in the model picker. Without access, the public gallery lets you watch previews.

Get the next AI explainer by email.

Published
Category
AI
Related Articles
Gemini 4 Neon: What It Does, What It Costs and How to Start

Gemini 4 Neon: What It Does, What It Costs and How to Start

Where Gemini 4 Neon fits beside Carbon, Hydrogen and Argon, who can use it, and how to compare it with your current model on a real coding task.Nov 19, 2026AI
Always-On Agents: Anthropic's Five Primitives, Shown Working

Always-On Agents: Anthropic's Five Primitives, Shown Working

An always-on agent mostly sleeps. Anthropic's five primitives wake it, carry its memory, guard its actions and deliver its work. Costs and limits included.Nov 19, 2026AI
Claude Cowork Pricing: What One Seat or a Team Costs

Claude Cowork Pricing: What One Seat or a Team Costs

Which Claude plans include Cowork? Compare US prices, usage limits, and worked budgets for one person, five seats, or twenty seats.Nov 19, 2026AI
QuillBot Review

QuillBot Review

When does QuillBot earn a seat beside your AI assistant? Compare its editing tools, US prices, team controls and privacy terms for client work.Nov 19, 2026AI
GPT-6.2 Astra vs Sol

GPT-6.2 Astra vs Sol

Choose a GPT-6.2 default by job. Compare Astra and Sol for agent runs, code review and writing, with worked examples and Ultrafast access by plan.Nov 18, 2026AI
GPT-6.2 Explained: Astra, Sol and Who Gets Access

GPT-6.2 Explained: Astra, Sol and Who Gets Access

Find GPT-6.2 Astra and Sol in ChatGPT Work, Codex and the API. See who gets Ultrafast and which jobs justify switching models.Nov 18, 2026AI
Kimi vs Qwen

Kimi vs Qwen

Kimi K3.1 Pro vs Qwen4-Max for coding agents and products: current access, long context, open weights, API tradeoffs and US deployment choices.Nov 17, 2026AI
Qwen 4: Qwen4-Max Explained

Qwen 4: Qwen4-Max Explained

Qwen4-Max explained: production versus preview access, context and output limits, open-model promises, and when to move from Qwen3.8-Max.Nov 17, 2026AI
Newsletter

One letter, every Sunday.Working systems, not hot takes.

Weekly. No spam. Unsubscribe anytime.