Google DeepMind has officially unveiled Genie 4, its world artificial intelligence (AI) model that introduces real-time, interactive world simulation with native audio and realism.
Alphabet and Google CEO Sundar Pichai highlighted the milestone, calling Genie 4 “our most realistic and interactive world model.” The system allows users to generate and actively navigate dynamic environments, steering scenes on the fly while the model simultaneously calculates physics, character behavior, and soundscapes.
Genie 4 is not a standalone graphics engine; it directly integrates the multimodal reasoning capabilities of Google’s Gemini family: Gemini Omni: Anchors each simulation in authentic real-world appearances, physical dynamics and visual continuity. Google says that Genie 4 delivers a significant leap in physical accuracy, accurately rendering complex interactions such as aerodynamics, collisions, and environmental resistance. – steer dynamic environments in real time, shaping events, camera angles, and actions – create more coherent multi-step events through reasoning from Gemini We’ve started to roll out as a research preview to trusted testers. Coming soon to Google AI Ultra subscribers in Project Genie and early access partners via the Gemini API and AI Studio. Character animation sees an equally sharp upgrade: digital figures move naturally, interact organically with their surroundings, and respond immediately to real-time steering.