🔍 Read the full analysis: Google’s Genie 4 Can Build Worlds—Here’s What To Expect on ThorstenMeyerAI.com
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Google DeepMind demonstrated Genie 4, an interactive AI world model that generates moving scenes and sound in response to user controls and text prompts. The system is not publicly available; DeepMind says scenes may drift and support only a few minutes of continuous interaction.
Google DeepMind demonstrated Genie 4 on October 27, presenting an AI model that generates interactive scenes with sound as users move through them or change them with text prompts. The system is not publicly available: Google says access is planned for trusted testers, with no public launch date, while also acknowledging that scenes can drift and continuous interaction lasts only a few minutes.
In DeepMind’s demonstration, a user could steer a generated scene with keyboard and mouse controls and enter instructions to alter it while it continued running. The company showed examples involving weather, objects, people and camera movement, alongside generated sounds such as wind, water and music. These demonstrations were selected and published by DeepMind, rather than independently tested results.
DeepMind describes Genie 4 as a real-time world model: it predicts what the next moment in a scene should look and sound like based on the preceding frames and user input. Unlike a conventional game engine, the model does not rely on a stored 3D map of the environment. That approach can produce detailed scenes, but DeepMind says objects and details can shift or drift over time.
The company says Gemini 3.8 Flash can interpret goals, break them into steps and help guide the generated scene. DeepMind also described audio generation and simulations involving physical events. Its robotics example showed a simulated Apptronik Apollo humanoid playing checkers; according to the source material, no physical robot took part in that demonstration. DeepMind has not published benchmark figures for the claimed capabilities.
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
What Interactive Worlds Could Enable
Genie 4 points toward AI systems that do more than produce a finished video: users can act inside a generated scene and alter it while it unfolds. If the approach becomes more reliable and broadly available, it could offer filmmakers and game developers a way to test settings or prototypes through interaction rather than building every environment by hand. DeepMind says testers have reported uses in film previsualization and game prototyping; those are early user reports, not evidence of established commercial results.
The model may also matter to robotics research, where simulated environments can help researchers explore tasks before testing them on machines. But the checkers example was a simulation, not a demonstration of a robot learning or completing a real-world task. Google’s existing use of the earlier Genie 3 model in a Waymo driving simulator provides relevant company context, but it does not establish that Genie 4 is ready for that role.
Generated scenes paired with sound may also make synthetic media more convincing. DeepMind says Genie 4’s video and audio outputs carry an invisible SynthID watermark. The source material does not describe how reliably that mark can be detected after content is edited or shared, so the claim should not be treated as a complete solution to identifying synthetic media.
Products Worth Considering
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
Logitech MX Mechanical Mini Mac Wireless Keyboard Low-Profile – Space Grey
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
As an affiliate, we earn on qualifying purchases.
How Genie Has Developed
Genie is part of a sequence of DeepMind world-model projects. The first Genie, announced in February 2024, generated simple 2D worlds from video data. Genie 2 followed in December 2024 with 3D environments created from a single image; DeepMind said those scenes could remain consistent for up to a minute, though its real-time version used reduced quality.
In August 2025, DeepMind introduced Genie 3 as its first real-time world model, describing output at 720p and 24 frames per second and scenes that could remain consistent for a few minutes. Google later made Genie 3 available to paying U.S. subscribers through Project Genie, according to the supplied source material. Genie 4’s announced changes include more realistic scenes, audio generated alongside visuals and text-based changes during interaction, but the company has not provided independent benchmark comparisons showing how much it improves on Genie 3.
A world model differs from a video generator that renders a requested clip for later viewing. It repeatedly predicts the next part of a scene as the user acts. It also differs from a standard game, whose environment is built from a defined map and assets. Genie’s frame-by-frame approach helps explain both its flexibility and its reported limitation: the scene may not remain stable as interaction continues.
“Details and objects “can shift or drift over time.””
— Google DeepMind, describing Genie 4’s limitations
Access, Runtime and Reliability
Genie 4 is currently described as available only to trusted testers. DeepMind says it is coming to Google AI Ultra subscribers through Project Genie and to early Gemini API partners, but it has announced no date or API price. The source material identifies Google AI Ultra as a $200-a-month tier; that price is not a Genie 4 access price and could change.
The length of usable sessions also needs clarification. DeepMind’s stated limit is “a few minutes” of continuous interaction. TechCrunch and Bloomberg, as cited by heise, have reported scenes of up to about two minutes, but that specific figure is not in DeepMind’s own description. No public test results establish how often worlds drift, how well prompts are followed across different scenes, or how much compute and latency the experience requires.
DeepMind’s examples are company-selected demonstrations, and the source material does not provide independent evaluations of the model. It is also unclear which features will be included for subscribers or API partners, what restrictions will apply, and how the SynthID watermark performs once generated media is modified.
The Next Access Milestones
The next stated step is access for trusted testers, followed by planned availability for Google AI Ultra subscribers through Project Genie and early Gemini API partners. DeepMind has not supplied a schedule, pricing details for API access, or a complete list of launch features.
Readers will be able to judge the announcement more fully when access expands beyond selected demonstrations. Independent testing can establish how long scenes remain coherent, how reliably text instructions change a world, and whether generated sound stays synchronized with events. Until then, Genie 4 is a company-demonstrated research and product development, not a broadly available tool with independently verified performance.
Key Questions
What is Google Genie 4?
Genie 4 is a Google DeepMind world model that generates a scene over time in response to user actions and text instructions, rather than producing only a finished video clip.
Can the public use Genie 4 now?
No public access is described. DeepMind says the system is available to trusted testers and is planned for Google AI Ultra subscribers through Project Genie and early Gemini API partners. No launch date has been given.
How long can a Genie 4 scene run?
DeepMind says Genie 4 supports a few minutes of continuous interaction and warns that objects and details may drift. Reports cited in the source material put scenes at about two minutes, but DeepMind did not give that exact figure in its own description.
Does Genie 4 control a real robot?
The robotics example described was a simulation of an Apptronik Apollo humanoid playing checkers. The source material says no physical robot was involved in that demonstration.
Does Genie 4 generate sound?
DeepMind says Genie 4 generates audio alongside its scenes, including environmental sounds, music and character speech. Those capabilities have been shown in the company’s demonstrations; independent performance tests were not supplied.
Source: ThorstenMeyerAI.com
