🔍 Read the full analysis: Where Google’s Genie 4 Shines—and Where Its Worlds Break Down on ThorstenMeyerAI.com
Get the latest gadgets 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 introduced Genie 4 on October 27, showing an interactive world model that responds to controls and text while generating visuals and sound. The demonstrations are company-selected, external access is not yet open, and DeepMind says scenes can drift after a few minutes of interaction.
Google DeepMind announced Genie 4 on October 27, demonstrating an AI model that generates an interactive scene as users move through it and issue text instructions. The company says it can create visuals and audio together, but the system is not yet publicly available, its demonstrations are company-selected, and DeepMind acknowledges that objects and details can drift during sessions lasting a few minutes.
Genie 4 is described by DeepMind as a real-time world model: rather than rendering a world from a stored 3D map like a conventional game engine, it predicts what the next frames should look and sound like based on what has already happened and on user input. In the company’s demonstration, a user can move through a scene with keyboard and mouse controls and type instructions to change elements such as weather, objects or camera movement.
DeepMind says Genie 4 generates ambient sound, music and dialogue alongside the visuals. It also describes more natural-looking people and simulations of physical events, including collisions and aerodynamics. These are vendor claims illustrated by its own selected clips; the announcement, as described in the source material, included no benchmark figures for comparing performance with other systems.
DeepMind says a Gemini model works alongside Genie 4 to interpret goals, break them into steps and monitor generated frames. One robotics demonstration showed a simulated Apptronik Apollo humanoid playing checkers, but no physical robot was involved. DeepMind has also cited tester feedback about film previsualization and game prototyping; those uses are not evidence of general commercial performance.
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 Clips
Genie 4 points toward AI systems that do more than produce a finished clip: a user can change a scene while it runs and respond to what appears. If the approach becomes reliable and accessible, it could offer filmmakers a way to explore rough scene ideas and game developers a tool for trying out interactive prototypes without first building every environment in a conventional engine.
Those possibilities remain prospective. The currently described examples come from DeepMind’s demonstrations and tester reports, while the model’s short session duration and visual drift limit what can be inferred about sustained use. The announcement also has relevance beyond entertainment: synthetic worlds may be useful in research and simulation, but a generated scene should not be mistaken for a validated substitute for a real-world test.
Audio generation adds a separate public-facing concern. DeepMind says Genie 4 audio and video carry an invisible SynthID watermark. That is the company’s stated provenance measure; the announcement does not establish how reliably the watermark can be detected after content is altered or shared through other services.
Products Worth Considering
Logitech MX Mechanical Wireless Illuminated Keyboard Tactile (Graphite)
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
Logitech MX Keys S Wireless Keyboard, Graphite
As an affiliate, we earn on qualifying purchases.
From Genie 3 to Genie 4
Genie 4 follows a series of systems that DeepMind has presented as increasingly interactive. The first Genie, introduced in February 2024, created simple 2D worlds. Genie 2 followed in December 2024 with 3D environments generated from a single image, though its consistency was described as lasting up to a minute and its real-time version used reduced quality.
In August 2025, Genie 3 was presented as a real-time model running at 720p and 24 frames per second, with consistency lasting a few minutes and visual memory reaching back about a minute. Google later offered Genie 3 through Project Genie to paying US subscribers. TechCrunch reported a 60-second limit per world, partly related to compute costs; that reported limit is distinct from DeepMind’s own description of Genie 3’s capabilities.
Genie 4’s planned distribution is also limited for now. The source says Google expects to offer it to trusted testers, with access “coming soon” for Google AI Ultra subscribers through Project Genie and for early Gemini API partners. No launch date or API price has been specified.
“Details and objects “can shift or drift over time.””
— Google DeepMind
Access, Consistency and Evidence
It remains unclear when ordinary users will be able to try Genie 4. DeepMind has described access for trusted testers and said availability for certain subscribers and API partners is “coming soon,” but has given no public release date or API pricing.
The length and quality of usable sessions are also not fully established. DeepMind says interaction can last a few minutes, while TechCrunch and Bloomberg, as cited by heise, reported scenes of up to about two minutes. That duration is not stated in DeepMind’s own text. The company has not provided public benchmark results or independent evaluations in the source material, and its selected clips cannot establish how often scenes drift or how well the model performs across varied prompts.
Likewise, reports of use for game development and film planning reflect tester feedback, while the Apollo example was a simulation. The announcement does not show that Genie 4 has been validated for training physical robots or for dependable production workflows.
A Wider Tester Release Ahead
The next stated step is access for trusted testers, followed by planned availability through Project Genie for Google AI Ultra subscribers and through the Gemini API for early partners. Google has not provided dates, pricing or technical access terms, so the rollout schedule remains open.
More independent testing would help clarify how long scenes remain coherent, how consistently text instructions change a world, and whether generated sound stays synchronized with the visuals. Until those details are available, Genie 4’s demonstrated capabilities should be read as company-presented results, not as a public performance record.
Key Questions
What is Google Genie 4?
Genie 4 is a real-time world model from Google DeepMind. It generates the next moments of a scene in response to user movement and text instructions rather than relying on a stored 3D map, according to the company.
Can the public use Genie 4 now?
No public release is described. DeepMind says access is planned for trusted testers, with later availability expected for Google AI Ultra subscribers through Project Genie and early Gemini API partners. No date has been announced.
What are Genie 4’s main limitations?
DeepMind says objects and details can shift or drift over time and describes continuous interaction lasting a few minutes. Public benchmarks and independent evaluations are not provided in the source material.
Does Genie 4 train or control a real robot?
The cited Apollo checkers demonstration was a simulation, not a test involving a physical robot. The material does not establish that Genie 4 is ready to train real-world robots.
Does Genie 4 generate sound?
DeepMind says the model generates audio alongside video, including ambient sounds, music and dialogue. The company also says its video and audio outputs carry an invisible SynthID watermark.
Source: ThorstenMeyerAI.com
