🔍 Read the full analysis: How Google Genie 4 Works, What It Can Make, And Who Gets It on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Google DeepMind introduced Genie 4 on October 27, describing it as a real-time AI model that generates interactive scenes from user inputs. The company says it adds synchronized audio and live text changes, but access is limited to trusted testers, and the model can lose consistency after a few minutes.
Google DeepMind announced Genie 4 on October 27, a so-called world model that generates scenes users can move through and alter with keyboard, mouse and text inputs. The company says it can create matching audio in real time, but the system is not broadly available: access is limited to trusted testers, and DeepMind says scenes can drift during a few minutes of interaction.
Unlike a conventional video generator that renders a clip for viewers to watch, Genie 4 predicts what happens next as a user interacts with a scene. A person might move through a generated environment, then type an instruction such as changing the weather. DeepMind says the model updates the scene in response. It does not use a stored 3D map like a conventional game engine; instead, it draws on what it has generated so far to predict subsequent frames.
DeepMind describes the system as supporting keyboard and mouse controls, text-directed changes, more realistic people and physics, and audio generated alongside the visuals. The company says that audio can include ambient sounds, music and dialogue. It also says Gemini 3.8 Flash can monitor a scene and break a goal into steps while Genie 4 produces new frames. These capabilities are presented in company-selected demonstrations; DeepMind has not published benchmark figures in the source material.
Access remains restricted. DeepMind says Genie 4 is available to trusted testers and is coming to Google AI Ultra subscribers through Project Genie, as well as early Gemini API partners. The source describes the subscription tier as costing $200 a month, but gives no launch date for broader access or price for API use. TechCrunch and Bloomberg, as cited by heise, report scenes lasting up to about two minutes; that figure does not appear in DeepMind’s own description.
Infographic 01 · How Genie grew up
From flat sketches to worlds with sound, in under three years.
Each step as Google DeepMind described it. Session limits and consistency are different measures, so read the labels.
- Feb 2024
Genie 1Simple 2D worlds11 billion parameters, learned from unlabelled internet videos - Dec 2024
Genie 23D worlds from a single imageConsistent “for up to a minute”; real time only as a slimmed-down, lower-quality version - Aug 2025
Genie 3First real-time Genie720p at 24 frames per second, consistent “for a few minutes”, memory about one minute back - Jan 2026
Project GenieUS AI Ultra subscribers can play60 seconds per world, partly because of compute costs - May 2026Project Genie goes globalIn the $200 AI Ultra tier, for users 18 and older
- 27 Oct 2026
Genie 4Sound, live steering, more realismResearch preview for trusted testers; Ultra subscribers and API partners “coming soon”
Genie 3 vs. Genie 4
As described by Google DeepMind. Genie 4 column = vendor claims unless marked.
| Feature | Genie 3 (Aug 2025) | Genie 4 (Oct 2026) |
|---|---|---|
| Sound | Genie 3: None | Genie 4: Generated with the picture: ambient noise, music, dialogue |
| Steering | Genie 3: Keys, plus “promptable world events” | Genie 4: Free-form text changes while it runs; Gemini adapts and checks them |
| Look | Genie 3: 720p, 24 fps | Genie 4: “Unprecedented realism”; no resolution or frame rate published |
| How long | Genie 3: “A few minutes”; Project Genie capped at 60 s | Genie 4: “A few minutes” (DeepMind); up to ~2 min consistent scene reported |
| Access | Genie 3: Small preview, then Project Genie | Genie 4: Trusted testers; Ultra and API partners “coming soon” |
Sources: Genie paper, arXiv, 23 Feb 2024; Genie 2 blog, 4 Dec 2024; Genie 3 blog, 5 Aug 2025; TechCrunch, 29 Jan 2026; Google, 19 May 2026; Google DeepMind, Genie 4 blog, 27 Oct 2026; Genie 4 model page (limitations, disclaimer); TechCrunch, 27 Oct 2026. Table wording: Thorsten Meyer.
Powered by Thorsten Meyer AI
Why Interactive Scenes Matter
If the demonstrations hold up beyond selected clips, Genie 4 could give creators and researchers a way to explore generated environments rather than simply watch finished videos. DeepMind says early testers have used the system for film previsualization and game prototypes. Those are reported tester uses, not independently measured productivity gains or a guarantee that the tool is ready for production work.
The technology also has a possible role in robotics research. In a demonstration described by DeepMind, Gemini 4 Argon planned steps for a task while Genie 4 simulated an Apptronik Apollo humanoid playing checkers. The source material says no physical robot was involved in that example, so it should be understood as a simulation rather than evidence of a robot performing the task in the real world. Separately, Waymo is reported to use Genie 3, an earlier model, in its driving simulator.
Genie 4’s ability to generate both visuals and sound also adds to the difficulty of judging synthetic media from a short clip alone. Google says every Genie 4 video and audio output carries an invisible SynthID watermark. The source does not explain how reliably that watermark can be detected after editing or redistribution.
Popular right now
DeLonghi Rivelia Automatic Espresso Machine Bundle with Grinder and Frother, Beige
Dreame L50 Ultra Robot Vacuum and Mop
Vitamix 5200 Professional-Grade Blender, 64 oz. Container, Black
As an affiliate, we earn on qualifying purchases.
From Genie Research to Live Worlds
Genie 4 follows a series of DeepMind world-model releases. The first Genie, introduced in February 2024, generated simple 2D worlds; its research paper described an 11-billion-parameter model trained on unlabeled internet videos. Genie 2, announced in December 2024, generated 3D worlds from a single image and kept them consistent for up to a minute, though real-time use required a reduced-quality version.
In August 2025, DeepMind presented Genie 3 as its first world model to run in real time, at 720p and 24 frames per second, with consistency lasting a few minutes. The source says Google opened Genie 3 through Project Genie to paying US subscribers in January 2026, with a reported 60-second limit per world, partly linked to computing costs. That history helps explain Genie 4’s focus on longer interaction and greater realism, but it does not establish that the new system has solved the earlier models’ limits.
“Details and objects “can shift or drift over time.””
— Google DeepMind, describing Genie 4’s limitations
Limits and Access Still Unsettled
Genie 4 has not been described with public benchmark results in the source material, and the demonstrations were selected by DeepMind. It is not yet clear how consistently the model performs across different prompts, how much control users will have over scenes, or whether its output quality will match the examples in routine use. DeepMind itself acknowledges that details and objects may drift over time.
The precise maximum length of a session is also unsettled. DeepMind says interaction can continue for a few minutes, while TechCrunch and Bloomberg, as cited by heise, report clips or scenes of up to about two minutes. The source does not provide a firm limit from Google. It also gives no date for access through Project Genie or the Gemini API, no API pricing and no details about how the SynthID watermark works in practice.
What Early Access May Show
The next concrete milestone is access for Google AI Ultra subscribers through Project Genie and for early Gemini API partners. Google has not announced when either group will receive access, how many users will be included, or what API access will cost. Until those details are published, Genie 4 remains a limited-access system rather than a generally available product.
Feedback from testers and any further technical documentation will help clarify whether the model can maintain stable scenes, respond reliably to instructions and support the proposed creative and research uses. For now, the confirmed picture is narrower: DeepMind has announced an interactive world model and shown selected demonstrations, while public evidence about performance outside those demonstrations remains limited.
Key Questions
What is Google Genie 4?
Genie 4 is an AI world model from Google DeepMind. It generates a scene frame by frame and responds as a user moves through it or gives instructions, rather than producing only a fixed video clip.
Can Genie 4 generate sound?
DeepMind says it generates audio alongside the visuals, including environmental sounds, music and dialogue. That is the company’s description; the source does not provide independent performance tests.
Who can use Genie 4 now?
Access is currently described as limited to trusted testers. DeepMind says Google AI Ultra subscribers through Project Genie and early Gemini API partners will get access soon, but it has not announced a date or API price.
How long can users interact with a Genie 4 scene?
DeepMind says Genie 4 supports a few minutes of continuous interaction. Media reports cited in the source put scenes at about two minutes, but DeepMind has not stated that precise limit in its own description.
Is Genie 4 ready for robotics or commercial production?
The source describes film previsualization and game prototypes as tester feedback, and a robotics example as a simulation. It does not establish that Genie 4 is ready for commercial production or that a real robot performed the demonstrated task.
Source: ThorstenMeyerAI.com
