Google DeepMind officially released its new world model Genie4 on October 27, 2026, marking a new stage in real-time interactive simulation technology.
This model deeply integrates unprecedented realism and native audio into real-time interactive simulation. By integrating the core capabilities of the Gemini series, it anchors simulations to the real world's appearance and behavior with Gemini Omni, while relying on other Gemini models for complex reasoning support to ensure the coherence and continuity of the simulation.
In terms of technical functionality, Genie4 allows users to launch events and control scenes in real time, including dynamically moving the camera, changing weather conditions, or introducing entirely new characters. Each simulated world can instantly generate native audio, including ambient sound effects, background music, and interactive dialogue, greatly enhancing the immersion and realism of the interaction.
As a key milestone on the path to artificial general intelligence, this model can not only provide safe virtual training environments for robots and agents, but also bring efficient innovative tools to diverse industries such as film production, news media, game development, and manufacturing.
Currently, Genie4 has been opened to some trusted testers and will soon be gradually rolled out to Google AI Ultra subscribers (through Project Genie) as well as early access partners of the Gemini API and AI Studio, further promoting the commercialization and ecosystem adoption of cutting-edge simulation technology.