Google announces Genie 4 world model with audio and real-time prompts
Google DeepMind has released Genie 4, a world model for interactive AI-generated worlds, as a research preview for trusted testers. It adds audio, more realistic scenes and mid-simulation text prompts, and runs for up to two minutes.
1 / 3
The story, neutrally told
Mixed · 2Google announced Genie 4, a world model that generates interactive experiences using AI, on Tuesday 27 October 2026. The VergeLC “Google announced today its Genie 4 world model for generating interactive experiences using AI.” Read at The Verge ↗ TechCrunchN “Google DeepMind released its newest world model, Genie 4, in research preview on Tuesday.” Read at TechCrunch ↗ Mixed · 2Google says the update brings more realistic worlds, the ability to change a scene in real time with a natural-language prompt, and generated audio. TechCrunch describes the audio as music, ambient noises and interactive dialogue that match the simulated world. The VergeLC “lets users update an experience in real time with a natural language prompt, can generate audio” Read at The Verge ↗ TechCrunchN “including music, ambient noises, and interactive dialogue.” Read at TechCrunch ↗ Mixed · 2Experiences can now last about two minutes, up from one minute for Genie 3. TechCrunch adds that edits can include changing the weather or camera angle or adding people or objects, and that physics and characters are improved. The VergeLC “It also can create experiences that are about two minutes long, up from one minute for the previous Genie 3 model.” Read at The Verge ↗ TechCrunchN “Genie 4 allows for freeform edits, like changing the weather or camera angle in real time or adding new people or objects.” Read at TechCrunch ↗
Mixed · 2The Verge reports that Genie 4 uses Gemini 3.8 Flash to help with its simulations. TechCrunch says Gemini acts as a director that plans steps and checks frames while Genie creates the visuals. The VergeLC “uses Gemini 3.8 Flash under the hood to help with its simulations” Read at The Verge ↗ TechCrunchN “acting as the director that plans steps and checks frames while Genie creates the visuals” Read at TechCrunch ↗ Centre · 1In a research experiment tied to Gemini Robotics, a model TechCrunch calls Gemini 4 Argon broke tasks into steps that Genie 4 simulated in real time. Google suggests this could allow robot scenarios to be tested in simulation first. TechCrunchN “a Gemini 4 Argon model broke down tasks into steps that Genie 4 simulated in real time” Read at TechCrunch ↗ Mixed · 2Google pitches Genie 4 for film pre-visualization, game prototyping, interactive learning and robot training scenarios. Google DeepMind engineer Diego Rivas said real-time generation is most valuable for ideation and storyboarding, and testers are already pointing to film prototyping. The VergeLC “pre-visualization for filmmakers, prototyping game ideas, and making training scenarios for robots” Read at The Verge ↗ TechCrunchN “The real time aspect becomes most valuable in the areas where keeping a state of flow in the creativity space becomes incredibly useful” Read at TechCrunch ↗
Mixed · 2Google acknowledges limits. Details and object permanence can shift as simulations run longer, and TechCrunch reports the team has not yet worked out how to ensure accurate audio and legible in-world text. Rivas stressed it is early days and the end-to-end product experience is unclear. The VergeLC “as simulations run longer, details and object permanence can shift over time” Read at The Verge ↗ TechCrunchN “they still haven’t quite figured out how to ensure the audio will be accurate and any in-world text will be legible” Read at TechCrunch ↗ Mixed · 2Genie 4 is available first as a research preview for trusted testers. It is coming soon to Google AI Ultra subscribers in Project Genie, and a waitlist is open. TechCrunch says early access partners of the Gemini API and Google AI Studio will also get it in the coming weeks. The VergeLC “Genie 4 will be “coming soon” to Google AI Ultra subscribers” Read at The Verge ↗ TechCrunchN “early access partners of the Gemini API and Google AI Studio” Read at TechCrunch ↗ Mixed · 2TechCrunch places the release in a crowded and consolidating field. Runway released GWM Worlds 2 a couple of months ago, World Labs is being acquired by AMD, Amazon is reportedly in talks to buy Decart, and Rivian has acquired Third Dimension. The Verge's writer has not tried Genie 4 and could not compare it with Genie 3-powered Project Genie. TechCrunchN “Fei-Fei Li’s World Labs is being acquired by AMD, Amazon is in talks to buy Decart, and Rivian just acquired Third Dimension” Read at TechCrunch ↗ The VergeLC “I haven’t been able to try Genie 4 for myself” Read at The Verge ↗
Every sentence links to the reporting it rests on. The pill in front of each says where its sources sit: Left, Centre or Right when one side supplies at least half of them, Mixed when they are evenly split. The number is how many outlets it cites.
Left2 outlets
- Framing
- Short product-news item built around Google's announcement and its four main updates, with the writer noting they have not tried the model.
- Emphasis
- The list of updates, the two-minute length versus Genie 3, Google's stated use cases and the limited rollout.
- Leaves out or plays down
- Does not cover the competitive landscape, the robotics experiment or the Google engineers' comments.
- Charged language
- “bad Nintendo knockoffs”
- For example
-
“I haven’t been able to try Genie 4 for myself — it’s launching first as a “research preview” to “trusted testers”” — The Verge
Centre1 outlet
- Framing
- Longer explainer on the release placed in the competitive, consolidating world-model industry, drawing on Google's press briefing.
- Emphasis
- Audio, steering mid-simulation, Gemini-Genie hybrid models, robotics experiment, industry deals and Google's admitted gaps.
- Leaves out or plays down
- Does not mention the specific Gemini 3.8 Flash model named by The Verge.
- Charged language
- “most photorealistic and interactive worlds yet”“stay ahead in the crowded field”
- For example
-
“Genie 4 is Google’s latest move to stay ahead in the crowded field.” — TechCrunch
“What we’ll increasingly see in the future are these hybrid world models” — TechCrunch
Right0 outlets
No right outlet in our sources has covered this story yet.
What every side reports
- Genie 4 was announced on 27 October 2026 as a research preview for trusted testers.
- It supports real-time prompt changes, audio and experiences of about two minutes, up from one minute for Genie 3.
- It is coming to Google AI Ultra subscribers through Project Genie.
- Details and objects can drift as simulations run longer.
Google DeepMind organisation
Presents Genie 4 as a step toward hybrid Gemini-Genie world models useful for creative prototyping and robotics, while stressing it is early and has gaps.
“Throughout the briefing, Rivas stressed that it’s still early days for Genie 4” — TechCrunch
Left2 articles
-
Google’s next Genie AI model generates more realistic experiences
Neutral Brief, factual rundown of Google's announcement, with a mildly sceptical note that the reporter could not test it.
-
Centre1 article
-
Google’s Genie 4 world model adds audio, more realistic worlds, and real-time steering
Neutral Detailed report on the launch that sets Genie 4 against industry competition and records Google's caveats.
Right0 articles
No coverage yet.
- 27 Oct 16:00 First The VergeLC Google’s next Genie AI model generates more realistic experiences
- 27 Oct 16:00 First Ars TechnicaLC DeepMind unveils Genie 4 world model promising more realistic physics and humans
- 27 Oct 16:30 +31m TechCrunchN Google’s Genie 4 world model adds audio, more realistic worlds, and real-time steering
Times are when each article was published, or when we first saw it if the outlet gave no time.
Ask about this story
Loading…