Genie 4: a new generation of world simulation
Google DeepMind presents Genie 4, a real-time interactive world model that also generates audio, with more realistic physics and human movement. It can be steered by text and is in research preview for trusted testers, with broader access planned.
What the article says
Google DeepMind introduced Genie 4 on October 27, 2026, a world model for real-time interactive simulation. It builds on the Gemini family, which anchors scenes in real-world appearance and behavior and keeps them coherent.
The model natively generates visuals and audio (music, ambient sound, dialogue) and lets users change a running simulation with natural language, such as weather, camera, or adding a person or object. It simulates physical phenomena and human motion, and a Gemini agent plans goals step by step. Genie 4 is a research preview for trusted testers; access is announced as coming soon for Google AI Ultra subscribers through Project Genie and for early partners through the Gemini API. Google lists limitations, including visual drift in longer sessions, and says generated content carries a SynthID watermark.
Read the official article AnnuIA profile
Summary written by AnnuIA in its own words from the official page of Google DeepMind. The content and rights belong to the publisher: please read the original article.