Google DeepMind announces Genie 4, a real-time interactive world model with native audio

Archive entry: this covers an event from October 27, 2026, and we published it on November 8, 2026. The details come from sources dated at the time and may have changed since.

On October 27, 2026, Google DeepMind announced Genie 4, a research world model for real-time interactive simulation with native audio, in a post on Google's The Keyword blog. At announcement it was available to limited trusted testers.

What Genie 4 does

• It builds on Gemini Omni and other Gemini models.
• Users can steer the simulation with natural language, including the camera, weather and characters.
• Its native audio covers ambient sound, music and dialogue.
• Google said it was coming to Google AI Ultra subscribers in Project Genie and to early-access partners in the Gemini API and AI Studio.

What it means for people building with AI

For developers, the access plan is the notable detail: Genie 4 was announced as coming to early-access partners in the Gemini API and AI Studio, which means Google's interactive world-model line was moving toward API access.

The capabilities Google listed describe a model that takes natural-language direction over the camera, weather and characters while producing its own ambient sound, music and dialogue. Google called Genie 4 a research model, and the wider access it described was still to come at announcement.

Other sources

Genie 4 also appeared in the Google DeepMind blog listing as a post about world simulation.

Sources

• We're announcing Genie 4, our new world model for real-time interactive simulation (Google (The Keyword), October 27, 2026, official)
• Google DeepMind blog listing: Genie 4: A new generation of world simulation (Google DeepMind, official)

Written with AI assistance, strictly from the sources listed above. Before it went live, automated checks verified its dates, sources and every language edition.