🔍 Read the full analysis: Exploring Genie 4: How Google Builds Worlds You Can Walk Into 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 on Oct. 27, describing it as a real-time world model that generates interactive scenes with sound. The company says access is coming to selected testers and partners, but it has not announced a public release date or published benchmark figures; DeepMind also says scenes can drift and support only a few minutes of interaction.
Google DeepMind demonstrated Genie 4 on Oct. 27, presenting it as a real-time AI model that generates a scene as a user moves through it and gives instructions. The company says it can create visuals and audio together, but the system is not yet broadly available, and DeepMind acknowledges that objects and details can drift during use.
Genie 4 is described by DeepMind as a world model: rather than rendering a scene from a stored 3D map, it predicts what should appear next from previous frames and user inputs. In the company’s demonstrations, users move through scenes with keyboard and mouse controls and enter text instructions to change details such as weather or objects. DeepMind says the model can generate environmental sounds, music and dialogue alongside the imagery.
DeepMind also says Gemini 3.8 Flash can interpret a user’s goal, break it into steps and monitor the changing scene. Its published demonstrations include a robotics-related simulation in which a humanoid robot plays checkers. The source account says this was a simulation, not a test involving a physical robot. DeepMind has not published benchmark figures in the material provided, and its demonstrations are company-selected examples rather than independent evaluations.
Access is currently described as limited to trusted testers. DeepMind says access is coming to Google AI Ultra subscribers through Project Genie and to early Gemini API partners, but it has given no launch date or API price. The source account describes Google AI Ultra as a $200-per-month subscription tier; that price and plan status should be checked against current Google materials before publication.
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 Move Beyond Video
If the demonstrations translate into a usable product, Genie 4 could give developers a way to explore scene ideas by interacting with them rather than watching a finished clip. DeepMind has reported tester feedback involving film previsualization and game prototypes, while its robotics example is a research demonstration. Those uses remain early and should not be treated as established commercial outcomes.
The combination of generated images and sound also makes provenance relevant. The source says Google will apply an invisible SynthID watermark to Genie 4 video and audio outputs. Watermarking may help identify generated material, but the supplied information does not establish how reliably it can be detected or how it performs after editing or reposting.
For companies, the immediate development is not a widely available production tool but a restricted system whose capabilities are still being tested. Its practical value will depend on factors not yet quantified publicly, including consistency over time, operating cost, access terms and whether users can export or reuse generated scenes.
Products Worth Considering
Logitech MX Mechanical Full-Size Wireless Mechanical Keyboard, Tactile Quiet Switches, Backlit Keys, Bluetooth, USB-C, Multi-Device, Graphit
Logitech MX Mechanical Mini Mac Wireless Keyboard Low-Profile – Space Grey
Logitech MX Keys S Wireless Keyboard, Graphite
As an affiliate, we earn on qualifying purchases.
Genie’s Progress Toward Real-Time Worlds
Genie 4 follows a series of models that DeepMind has described as increasingly interactive. The source account says the first Genie, introduced in February 2024, generated simple two-dimensional worlds from online video. Genie 2, presented in December 2024, generated three-dimensional scenes from a single image and maintained them for up to about a minute, though real-time use was limited to a lower-quality version.
According to that account, Genie 3 arrived in August 2025 with real-time generation at 720p and 24 frames per second, maintaining a scene for a few minutes and retaining visual memory reaching back about a minute. Project Genie later gave paying US subscribers access to Genie 3, with a reported 60-second limit per world. These details provide a comparison point, but the supplied source does not include the underlying Google documentation for every figure.
Genie 4’s stated advance is the combination of real-time interaction, more realistic scenes and native audio generation. That description comes from DeepMind; the material provided does not include independent testing that measures how much the system improves on its predecessor.
“Details and objects can shift or drift over time.”
— Google DeepMind
Access, Performance and Limits
Several points remain unsettled. Google has not provided a public release date, API pricing or benchmark results in the material supplied. DeepMind says interaction lasts a few minutes, while reports cited by the source account describe scenes of about two minutes; the exact limit may depend on the version or test, but that is not explained.
It is also unclear how often scene drift occurs, how well instructions work across different settings, or how performance compares with other systems under the same conditions. The selected demonstrations cannot answer those questions on their own. The reported uses in film and game development are tester feedback, and the robotics example is a simulation rather than evidence of performance on a real robot.
Waiting for Wider Tester Access
The next confirmed step is limited access for trusted testers, followed by planned access for Google AI Ultra subscribers through Project Genie and early Gemini API partners. Google has not specified when those groups will be admitted or whether access will expand beyond them.
Further details to watch for include an official release schedule, API terms, independent evaluations and clearer limits on session length and scene consistency. Until those appear, Genie 4 is best understood as a demonstrated research and development system, not a generally available tool with established performance in commercial workflows.
Key Questions
What is Genie 4?
Genie 4 is a Google DeepMind world model that predicts and generates a scene as users move through it and provide inputs. DeepMind says it can generate audio alongside the visuals.
Can the public use Genie 4 now?
No public release is specified in the source material. DeepMind says access is limited to trusted testers, with broader access planned for Google AI Ultra subscribers through Project Genie and early Gemini API partners.
How long can a Genie 4 world run?
DeepMind describes support for a few minutes of continuous interaction. The source account cites reports of scenes lasting about two minutes, but says that duration is not stated in DeepMind’s own materials.
Does Genie 4 control a real robot?
The robotics example described in the source is a simulation involving a humanoid robot playing checkers. It does not establish that Genie 4 controlled a physical robot.
What limitations has DeepMind acknowledged?
DeepMind says objects and details can drift over time and describes sustained interaction as lasting only a few minutes. It has not supplied benchmark results in the source material.
Source: ThorstenMeyerAI.com
