🔍 Read the full analysis: From AI Prompt To Walkable World: A Guide To Google’s Genie 4 on ThorstenMeyerAI.com
Get the little things that make your day 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 on Oct. 27, presenting it as a real-time model that generates interactive scenes from video, user controls and text prompts. DeepMind says it can create matching audio and respond to changes, but also warns that scenes can drift after a few minutes. The system is not generally available; Google has not announced a release date or API pricing.
Google DeepMind demonstrated Genie 4 on Oct. 27, describing it as a real-time AI model that generates interactive scenes in response to movement and text instructions. The company says the system can produce video and audio together, but it remains limited to trusted testers, and DeepMind acknowledges that objects and details can drift during sessions lasting a few minutes.
In the demonstration, a user enters a scene description and moves through the resulting environment with keyboard or mouse controls. DeepMind says users can also type changes while the scene runs, such as altering the weather or adding an object. The company describes the model as generating each next moment from the scene so far and the user’s inputs, rather than rendering a world from a conventional, stored 3D map.
DeepMind lists more realistic people and physics, live text-driven changes and natively generated sound among Genie 4’s capabilities. The company says the model can produce ambient audio, music and dialogue alongside images. It also says Gemini 3.8 Flash acts as a guiding agent, interpreting goals, breaking them into steps and checking generated frames. These are company descriptions; DeepMind has not published benchmark figures in the supplied announcement material.
The system is not open to general users. DeepMind says access is planned for trusted testers, with a later release described as coming soon for Google AI Ultra subscribers through Project Genie and early Gemini API partners. The source material gives no release date or API price. DeepMind also says Genie 4 outputs carry an invisible SynthID watermark.
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 a way to make and explore virtual scenes without first building a conventional game environment. If the system works reliably beyond selected demonstrations, interactive generated worlds could help creators test visual ideas and let developers prototype scenes through prompts and controls. DeepMind has reported tester feedback involving film previsualization and game prototypes, but those examples are not evidence of broad commercial deployment or measured productivity gains.
The announcement also links world models to robotics research. DeepMind showed a simulation of Apptronik’s Apollo humanoid playing checkers, with a Gemini model breaking the task into steps. According to the source material, this was a simulation, not a test involving a physical robot. The distinction matters: a generated scene may offer a setting for research, but the demonstration alone does not show that a robot can perform the task in the real world.
For viewers, increasingly convincing scenes paired with sound may make it harder to judge how a clip was made from appearance alone. DeepMind says it applies SynthID to Genie 4 video and audio outputs. The watermark is a stated provenance measure, not proof that every generated clip will be identified by every viewer or platform.
Products Worth Considering
Sonos Era 100 SL Compact Microphone-Free Speaker, Black
Dyson Cyclone V10 Origin Cordless Vacuum
Bose TV Speaker (2nd Gen) Compact Soundbar with Bluetooth and HDMI-ARC, Black
As an affiliate, we earn on qualifying purchases.
Genie’s Move Toward Real Time
Genie 4 follows earlier versions that DeepMind presented as steps toward interactive generated environments. The source material says the original Genie, introduced in February 2024, created simple 2D worlds. Genie 2, announced in December 2024, generated 3D worlds from a single image and maintained them for up to a minute, while its real-time operation required a reduced-quality version.
DeepMind introduced Genie 3 in August 2025 as its first real-time world model, describing output at 720p and 24 frames per second and consistency lasting a few minutes. Google later made Genie 3 available to paying US subscribers through Project Genie in January 2026. TechCrunch reported a 60-second limit per world at that time, citing computing costs; that reported limit should not be treated as a published Genie 4 specification.
The new announcement continues that progression, with DeepMind emphasizing audio, live scene changes and assistance from Gemini. Its demonstration footage was selected by the company, so it shows what DeepMind chose to present rather than independent testing across a range of prompts and conditions.
“Details and objects can shift or drift over time.”
— Google DeepMind, describing Genie 4’s limitations
Access, Stability and Performance
Several practical details remain unanswered. DeepMind has not provided a public launch date, Gemini API pricing or independent benchmark results in the source material. It is also not clear how performance varies across prompts, scene types or longer sessions, or how well generated locations remain consistent when users turn away and return.
Published reports cited in the source material describe demonstration scenes lasting up to about two minutes, but that figure is attributed to TechCrunch and to Bloomberg as cited by heise; it is not stated in DeepMind’s own announcement text. The relationship between that reported duration and the company’s broader description of “a few minutes” is unclear. The selected demos also do not establish how often the model’s physics, characters or audio behave as described outside those examples.
When Testers Get Access
DeepMind says Genie 4 is planned for trusted testers, followed by access described as coming soon for Google AI Ultra subscribers through Project Genie and early Gemini API partners. The company has not given dates, eligibility details or pricing for API access, so the next concrete milestone is a formal announcement of availability and terms.
Until wider access or independent evaluations are available, the best evidence remains DeepMind’s selected demonstrations and its own description of the model’s limits. Future reporting can assess whether users can maintain coherent scenes for longer, how generated audio responds to changes, and whether the system proves useful for film, games or robotics beyond research examples.
Key Questions
What is Google Genie 4?
Genie 4 is Google DeepMind’s interactive world model. The company says it generates scenes as users move through them and enter text instructions, rather than relying on a conventional stored 3D map.
Can the public use Genie 4 now?
No public release is confirmed in the source material. DeepMind says access is planned for trusted testers, with later availability for Google AI Ultra subscribers through Project Genie and early Gemini API partners. No date has been announced.
Does Genie 4 generate sound?
DeepMind says it generates audio alongside video, including ambient sound, music and dialogue. These capabilities were described by the company and shown in its selected demonstrations; independent performance results are not provided.
How long can a Genie 4 world last?
DeepMind describes a few minutes of continuous interaction and warns that details or objects can drift. Reports of scenes lasting up to about two minutes are attributed to TechCrunch and Bloomberg as cited by heise, not to DeepMind’s announcement text.
Has Genie 4 been tested with a real robot?
The robotics example described in the source material was a simulation of an Apptronik Apollo humanoid playing checkers. It does not establish that a physical Apollo robot performed the task.
Source: ThorstenMeyerAI.com
