Google DeepMind introduced Genie 4 on October 27, an AI model designed to generate realistic virtual environments that users can interact with and modify in real time. The technology combines detailed visuals, physics-based interactions, reasoning and audio generation, allowing users to create and change simulated worlds using prompts. Google CEO Sundar Pichai described it as the company's "most realistic and interactive world model" in a post on X.
Google DeepMind Genie 4 begins research rollout #
Google DeepMind has begun testing Genie 4, its AI model for generating interactive virtual environments, with a select group of trusted testers. The company plans to expand access to Google AI Ultra subscribers through Project Genie, while early-access partners are expected to use the model through the Gemini API and Google AI Studio. The rollout is currently limited, with wider availability yet to be announced.
Excited to share Genie 4 today our most realistic and interactive world model: - steer dynamic environments in real time, shaping events, camera angles, and actions - generate native audio matched to the world, from realistic background sounds to characters who can converse in real time! - simulate photorealistic scenes and real-world physics - 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.
— Sundar Pichai (@sundarpichai)October 27, 2026
Change weather, move cameras and add characters with prompts #
Genie 4 lets users shape virtual scenes through natural-language instructions and see changes unfold in real time. Users can move the camera, alter weather conditions or introduce new characters without having to build every element manually. The model also uses reasoning capabilities from Google's Gemini family to help keep the simulated environment consistent as events progress.
Genie 4 combines realistic visuals, physics and audio #
Beyond generating detailed, lifelike scenes, Genie 4 can produce environmental sounds, background music and interactive conversations between virtual characters. Its aim is to make these worlds behave more like real environments, with physics-based interactions and connected sequences of events. In one demonstration highlighted by Google CEO Sundar Pichai, a prompt turns a traffic light green, prompting a motorcycle's engine to rev before the bike accelerates along a coastal road.
How Google's AI world model could be used #
Google DeepMind sees world models such as Genie 4 as a potential step towards artificial general intelligence (AGI). Beyond entertainment, these simulations could provide controlled environments in which robots and AI agents learn and practise tasks. The technology may also find applications in filmmaking, journalism, gaming and manufacturing, although its broader real-world use will depend on how the model develops and becomes available.
ALSO SEE: