🔍 Read the full analysis: Google’s Genie 4: A Hands-On Look At Its World-Building AI 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 demonstrated Genie 4 on Oct. 27, describing it as a real-time world model that generates interactive scenes and audio. The showcase uses company-selected clips; the model is not yet publicly available, and DeepMind says scenes can drift and support only a few minutes of interaction.
Google DeepMind demonstrated Genie 4 on Oct. 27, presenting an AI system that generates interactive scenes in response to keyboard, mouse and text instructions. The company says the model can create visuals and sound in real time, but the demonstrations were selected by DeepMind, public access has not opened, and the company acknowledges that scenes may drift after a few minutes.
Genie 4 is described by DeepMind as a world model: rather than rendering a finished video from a prompt, it predicts what happens next as a user moves through a scene or gives new instructions. In the company’s demonstrations, users steer with standard controls and type changes such as making it rain. DeepMind says the system generates environmental audio, music and dialogue alongside the imagery.
DeepMind also says Gemini 3.8 Flash can interpret goals, break them into steps and monitor the generated scene between user inputs. The company describes simulations of events such as collisions and aerodynamics. These are vendor claims, and the announcement material provided no benchmark results or independent evaluation data.
Access is limited to trusted testers for now. DeepMind says Genie 4 is planned for Google AI Ultra subscribers through Project Genie and for early Gemini API partners, but it has not announced a launch date or API pricing. The source reports that TechCrunch and Bloomberg, as cited by heise, described clips lasting up to about two minutes; that figure does not appear in DeepMind’s own wording, which says interaction can continue for “a few minutes.”
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
Genie 4’s importance lies in a shift from generating a clip to generating a scene that responds. If the system can maintain a coherent environment while people move through it and change it, it could offer a quicker way to sketch game settings, film sequences or simulations than building each scene with conventional tools.
DeepMind says testers have used the system for film previsualization and game prototypes; those examples are reported tester feedback, not evidence of broad commercial adoption. The robotics demonstration is also a simulation, not a robot operating in the physical world. The model’s short interaction window and drift may limit uses that depend on accuracy or consistent detail.
Generated scenes with matching audio also make it harder for viewers to judge whether a clip is synthetic. DeepMind says Genie 4’s video and audio outputs carry an invisible SynthID watermark. The material does not explain how reliably that watermark can be detected after editing or reposting.
Products Worth Considering
Logitech MX Mechanical Wireless Illuminated Keyboard Tactile (Graphite)
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
As an affiliate, we earn on qualifying purchases.
From Genie 3 to Genie 4
The announcement follows several stages of DeepMind’s Genie research. The first Genie, introduced in February 2024, generated simple 2D worlds. Genie 2, announced in December 2024, created 3D scenes from a single image and kept them consistent for up to a minute, according to the source material.
Genie 3, presented in August 2025, was described by DeepMind as its first real-time world model, running at 720p and 24 frames per second and maintaining consistency for a few minutes. Google later made it available to paying U.S. subscribers through Project Genie, according to the source. TechCrunch reported a 60-second limit per world, attributed partly to computing costs. Genie 4’s demonstration adds native audio and live text changes, but DeepMind has not supplied comparable performance figures.
““a few minutes of continuous interaction””
— Google DeepMind
Access, Consistency and Evidence
Genie 4 is not yet generally available, and DeepMind has not specified when testers outside its current group—or subscribers and API partners—will receive access. Pricing and usage limits for the Gemini API are also unannounced.
The company has not published benchmark numbers in the material provided, and the public examples are selected demonstrations rather than an independent test of performance. It remains unclear how well the model maintains locations, objects and physical behavior across longer sessions or more demanding tasks. The precise duration users can expect is also unsettled: DeepMind says “a few minutes,” while outside reports cited in the source give an approximate two-minute figure.
The robotics example does not establish that Genie 4 can safely or reliably train physical robots. The source says the checkers demonstration simulated an Apptronik Apollo humanoid; no real robot was involved.
Testing and Release Details
The next concrete steps are expected to be tester access and a wider rollout through Project Genie and early Gemini API partnerships, although Google has not provided dates. More information on subscription eligibility, API costs, technical limits and supported controls would help establish what users can actually do.
Independent access will also matter. Tests beyond DeepMind’s curated clips could show whether scenes remain coherent during longer interactions, how reliably audio matches events, and whether the system is useful for production or research tasks. Until those details emerge, Genie 4’s demonstrated capabilities should be treated as company-presented results, not a settled measure of real-world performance.
Key Questions
What is Google Genie 4?
Genie 4 is a Google DeepMind world model designed to generate interactive scenes as users move through them and issue instructions. DeepMind says it can produce visuals and audio together.
Can the public use Genie 4 now?
No. The source says it is available to trusted testers and is planned for Google AI Ultra subscribers through Project Genie and early Gemini API partners. No public release date has been announced.
How long can a Genie 4 world last?
DeepMind says it supports a few minutes of continuous interaction and warns that objects and details can drift. Outside reports cited in the source describe scenes of about two minutes, but the company has not stated that specific duration in its own description.
Can Genie 4 train real robots?
The cited robotics example was a simulation involving an Apptronik Apollo humanoid; the source says no physical robot took part. It does not establish that Genie 4 is ready for real-world robot training.
Source: ThorstenMeyerAI.com
