🔍 Read the full analysis: Could Google’s Genie 4 Change How We Explore AI-Generated Worlds? 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 presented Genie 4 on October 27, 2026, describing a real-time AI world model that responds to movement and text prompts while generating visuals and sound. The system is not publicly available, and DeepMind says scenes can drift and support only a few minutes of continuous interaction.
Google DeepMind presented Genie 4 on October 27, describing an AI system that generates an interactive world in response to keyboard, mouse and text inputs, with sound produced alongside the visuals. The model is not open to the public yet, and DeepMind says its scenes can drift over time and remain continuously interactive for only a few minutes.
DeepMind describes Genie 4 as a world model: rather than rendering a completed video from a prompt, it predicts what should appear next as a user moves through a scene or gives new instructions. The company says users can steer the generated environment with standard controls and make changes while it runs, such as adding rain, altering the camera or introducing an object.
The company also says Genie 4 generates audio in real time, including environmental sounds, music and character speech. DeepMind describes improved physical behavior, including collisions and aerodynamics, and more natural-looking people. Those descriptions come from Google’s selected demonstrations; the announcement included no benchmark figures to quantify the claimed improvements.
DeepMind says Gemini 3.8 Flash can interpret goals, divide them into steps and monitor the scene as it changes. In one research demonstration, a model called Gemini 4 Argon guided a simulated Apptronik Apollo humanoid through a game of checkers. The demonstration used a simulated robot, not a physical Apollo machine, according to the source report.
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
Interactive Worlds Beyond Video
If the system performs as described, Genie 4 could make generated environments more useful for tasks that require interaction, not just a finished clip. A designer might test a setting by moving through it, or change weather and objects without generating an entirely new scene. DeepMind has reported tester feedback describing uses in film previsualization and game prototyping; that feedback is not evidence of a released product or independently measured results.
For robotics, simulated environments can let researchers examine scenarios without running every test on physical hardware. The Apollo demonstration illustrates that research direction, but it does not show that Genie 4 has trained a real robot or improved its performance. Separately, the source says Waymo uses Genie 3, a predecessor, in its driving simulator; that does not establish that Waymo has adopted Genie 4.
Generating sound together with images may also make synthetic scenes feel more complete. That raises practical questions about identifying generated media. Google says Genie 4’s video and audio outputs carry an invisible SynthID watermark. The announcement does not provide details about how reliably that marker can be detected after editing or reposting.
Popular right now
Vitamix Propel Series 510 Professional-Grade Blender, 3 Programs, Black
De'Longhi Magnifica Evo Fully Automatic Espresso Machine with Manual Milk Frother, Silver
Vitamix Propel Series 750 Professional-Grade Blender, 4 Programs, Black
As an affiliate, we earn on qualifying purchases.
From Genie 3 to Genie 4
The Genie series has developed from research into increasingly interactive generated environments. The first Genie, introduced in February 2024, created simple 2D worlds. DeepMind’s December 2024 Genie 2 announcement described generating 3D worlds from a single image and maintaining consistency for up to a minute, while real-time use required a reduced-quality version.
In August 2025, DeepMind presented Genie 3 as its first real-time world model, reporting output at 720p and 24 frames per second and consistency for a few minutes. Google later made Genie 3 available to paying US subscribers through Project Genie, according to the supplied source. TechCrunch reported a 60-second limit per world in that service and attributed it partly to computing costs; that report concerns Genie 3, not Genie 4.
Genie 4’s proposed advance is the combination of interactive scene generation, live text changes and audio. DeepMind also says Gemini Omni anchors simulations in real-world appearance and behavior. These are company descriptions, and the available announcement material does not offer independent testing or numerical comparisons with earlier models.
““a few minutes of continuous interaction””
— Google DeepMind
Access, Reliability and Evidence
Genie 4 is currently limited to trusted testers. DeepMind says access is coming to Google AI Ultra subscribers through Project Genie and to early Gemini API partners, but has not announced a launch date or API price. The source identifies Google AI Ultra as a $200-a-month subscription tier; the terms and availability of any Genie 4 access were not specified.
It is not yet clear how Genie 4 performs outside the company’s selected demonstrations, how long a session can last in practice, or how consistently a scene preserves objects and layout. DeepMind’s description says interaction lasts a few minutes and admits that details can drift. TechCrunch and Bloomberg, as cited by German outlet heise, reportedly described scenes lasting up to about two minutes; that figure is not stated in DeepMind’s own announcement, so the exact limit remains uncertain.
The announcement also provides no benchmark results for realism, latency, physical accuracy or cost. The announced capabilities and proposed use cases therefore remain vendor claims or early tester feedback, not independently established performance. No detailed information is provided about SynthID’s detection performance or how the model handles user-created content.
Testing and Access Milestones
The next expected step is access for trusted testers, followed by the planned availability for eligible Google AI Ultra subscribers in Project Genie and early partners using the Gemini API. Google has not provided dates, pricing beyond the existing subscription reference, or a public API release schedule.
Further demonstrations or testing may clarify whether the model can keep environments stable, respond quickly to controls and generate useful audio over longer sessions. Readers should distinguish those future tests from what has been shown so far: company-selected demos and limited tester feedback. Until access expands and performance data is available, Genie 4’s practical uses and limits remain developing.
Key Questions
What is Google Genie 4?
Genie 4 is an AI world model that DeepMind says predicts and generates a scene as a user moves through it or provides new instructions. Unlike a conventional video generator, it is designed for ongoing interaction.
Can the public use Genie 4 now?
No. DeepMind says access is currently for trusted testers. The company has said it plans to offer access through Project Genie to Google AI Ultra subscribers and to early Gemini API partners, but has not announced a date.
What limits has DeepMind acknowledged?
DeepMind says objects and details can drift over time and describes continuous interaction lasting a few minutes. An exact session limit for Genie 4 has not been specified in the company’s announcement.
Does the robot demonstration show Genie 4 controlling a real robot?
No. The described Apptronik Apollo demonstration was a simulation. The supplied report says no physical robot was involved, so it does not establish real-world robot performance.
How does Google say people can identify Genie 4 media?
Google says video and audio generated by Genie 4 carry an invisible SynthID watermark. The announcement does not explain how well it can be detected after media is altered or redistributed.
Source: ThorstenMeyerAI.com
