The facts

  1. Google says users can steer scenes in real time with text, including changing weather, camera position or objects. Native audio includes ambient sound and dialogue. [1]

  2. The planned next audiences are Google AI Ultra subscribers through Project Genie and early-access partners using the Gemini API and Google AI Studio. The announcement is not general public availability. [1][2]

  3. Google identifies limits in object permanence, physical interactions, audio accuracy, readable text and long-session consistency. Its realism and simulation claims remain the maker’s claims. [1]

  4. TechCrunch reported from a company briefing that the model can maintain a scene for up to two minutes and that the eventual end-to-end product experience remains unsettled. [2]

Still unconfirmed

  • A general-release date, full regional availability and product-specific public pricing were not established in the checked announcement.

Sources

References support the numbered facts above. Source checks: 29 Oct 2026, 00:57 UTC.

  1. Genie 4 announcement and limitations

    Google DeepMind · Primary source · 27 Oct 2026

  2. Independent briefing report on Genie 4

    TechCrunch · Reporting · 27 Oct 2026

Report a factual error