Choose Rich Live
Tech

Google DeepMind unveils Genie 4 interactive AI world model

Google DeepMind has introduced Genie 4, an AI world model that creates interactive real-time simulations featuring native audio, realistic physics, and live prompt-based editing.

Graphic for Google DeepMind's Genie 4 displaying a spherical lens view of an interactive simulation featuring a crowned crane near a piano keyboard in a snowy landscape.
Image: @iPhoneinCanada

Google DeepMind has unveiled Genie 4, an AI world model capable of generating photorealistic, interactive 3D simulations in real time, complete with physics, audio, and dialogue. The model turns text prompts into explorable environments that users can navigate using a keyboard and modify on the fly by typing, allowing them to shift weather conditions, adjust camera angles, or introduce new objects and people.[2][7][8][9][11][12]

Genie 4 brings notable technical upgrades to its simulated environments, including the synthesis of real-time background noise, music, and spoken conversation, alongside more realistic human figures. The model also maintains scene consistency for up to two minutes, which is twice the duration supported by its predecessor, Genie 3.[1][4][8][12]

According to Jay Peters of The Verge, Genie 4 is powered by Gemini 3.8 Flash. DeepMind credited the announcement post to Raia Hadsell, Diego Rivas, and Hang Qi, releasing the system as a research preview for trusted testers alongside a waitlist, with plans to bring the technology to Google Ultra subscribers through Project Genie.[3][5][6][9][10]

Key facts

  • Google DeepMind unveiled Genie 4, a world model that transforms text prompts into real-time interactive 3D simulations.
  • Genie 4 generates real-time audio, including music, background sounds, and conversation dialogue.
  • Users can steer simulations using mid-simulation text prompts and navigate the generated worlds using a keyboard.
  • The model supports consistent scenes for up to two minutes, doubling the duration of Genie 3.
  • The Verge reported that Genie 4 is powered by Gemini 3.8 Flash.
  • Genie 4 is initially available as a research preview to trusted testers, with a public waitlist and planned access for Google Ultra subscribers.

Sources · 12 sources

  1. AT

    Ars Technica@arstechnicaPost on X ·

    DeepMind's Genie 4 world model adds audio and more realistic people to simulations https://t.co/sDvqaWzAVG

    Open source
  2. AG

    AGTP@AGTPinsightsPost on X ·

    BREAKING: Google DeepMind unveils Genie 4, an AI world model that turns text prompts into explorable, real-time worlds you can navigate with a keyboard and edit by typing. https://t.co/UMe4ruBlcs

    Open source
  3. TE

    TechmemeArticle ·

    Google launches Genie 4, a world model powered by Gemini 3.8 Flash designed to generate interactive AI simulations, initially as a preview to trusted testers (Jay Peters/The Verge) Jay Peters / The Verge : Google launches Genie 4, a world model powered by Gemini 3.8 Flash designed to generate interactive AI simulations, initially as a preview to trusted testers — The Genie 4 world model offers more realism and can create audio to accompany its interactive worlds.

    Open source
  4. CB

    Crypto BriefingArticle ·

    DeepMind’s Genie 4 brings audio and sharper realism to AI-generated worlds Genie 4's advancements in realism and audio could accelerate AI's role in diverse fields, impacting industries beyond gaming significantly. The post DeepMind’s Genie 4 brings audio and sharper realism to AI-generated worlds appeared first on Crypto Briefing .

    Open source
  5. LA

    Lisan al Gaib@scaling01Post on X ·

    Google just announced Genie 4 https://t.co/X6DiOL5htZ

    Open source
  6. TE

    Techmeme@TechmemePost on X ·

    Google launches Genie 4, a world model powered by Gemini 3.8 Flash designed to generate interactive AI simulations, initially as a preview to trusted testers (@jaypeters / The Verge) (Visit Techmeme dot com for the link and full context!)

    Open source
  7. CB

    Crypto Briefing@Crypto_BriefingPost on X ·

    🌎NEW: Google unveils Genie 4, its latest AI world model capable of generating photorealistic, interactive 3D environments with native audio and realistic physics. https://t.co/FDfT0UtGqc

    Open source
  8. PO

    Polymarket@PolymarketPost on X ·

    JUST IN: Google DeepMind unveils Genie 4, a new AI model that generates photorealistic, interactive 3D worlds in real time, complete with physics, sound, & human dialogue.

    Open source
  9. TE

    TechCrunch@TechCrunchPost on X ·

    Google DeepMind released Genie 4, a world model with real-time audio, more realistic simulations, and mid-simulation text prompts, in research preview for trusted testers. https://t.co/z7wmCMXK4h

    Open source
  10. MT

    MTS@MTSlivePost on X ·

    SITUATION DETECTED: Google has released their Genie 4 world model in research preview, and announced that it is coming soon to Ultra subscribers in Project Genie.

    Open source
  11. II

    iPhone in Canada@iPhoneinCanadaPost on X ·

    Google's Genie 4 Builds Live Worlds You Steer With Plain Words https://t.co/bqEL1wMThc https://t.co/dZTBIC8r66

    Open source
  12. TD

    The Daily Block@thedailyblockPost on X ·

    🚨 GOOGLE DEEPMIND LAUNCHES GENIE 4, ITS NEW AI WORLD MODEL Google has introduced Genie 4, which can create realistic, interactive worlds with sound. Users can change the weather, camera angles, or add new people and objects while exploring the AI-generated world. The model can also generate music, background sounds and conversations in real time. Genie 4 supports consistent scenes for up to 2 minutes, twice as long as Genie 3. The model is currently available as a research preview.

    Open source