Google's Genie 4 creates realistic worlds and allows real-time changes
What's the story
Google DeepMind has unveiled its latest world model, Genie 4. The new model is capable of generating highly realistic and interactive virtual environments. Users can even change the scene by typing prompts in real time during a simulation. The release comes just months after AI video start-up Runway introduced GWM Worlds 2, its second world model, with interactive video and audio generation capabilities.
Enhanced functionality
More photorealism and audio capabilities
Genie 4 is a major upgrade from its predecessor, Genie 3.
The new model offers more photorealism and audio capabilities, along with freeform editing options like real-time weather or camera angle changes using prompts.
It can keep a scene stable for up to two minutes, twice as long as the previous version, and provides improved physics and more realistic characters.
AI collaboration
Gemini AI now integrated with Genie 4
Google's AI model, Gemini, is now more closely integrated with Genie 4.
It acts as a director that plans steps and checks frames while Genie creates visuals.
In a research experiment related to Gemini Robotics, a Gemini 4 Argon model broke down tasks into steps that Genie 4 simulated in real time.
This could be used to test robot scenarios before they happen in real life.
Future prospects
Potential applications in gaming and more
The commercial use cases for Genie 4 are gaming, interactive learning, and robotics.
One potential application could be entering a painting of a historical scene and chatting with its subjects about the event depicted.
Google DeepMind engineer Diego Rivas said that the real-time aspect of Genie 4 would be most valuable in areas where maintaining a state of flow in creativity is crucial, such as during rapid ideation phases.
Access and improvements
Research preview available for trusted testers
Genie 4 is now available as a research preview for trusted testers.
In the coming weeks, it will be rolled out to Google AI Ultra subscribers in Project Genie and early access partners of the Gemini API and Google AI Studio.
However, Rivas stressed that it's still early days for Genie 4, with gaps like detail drift over time in simulations and challenges ensuring accurate audio or legible in-world text.
Twitter Post
Demo posted by Sundar Pichai
Excited to share Genie 4 today our most realistic and interactive world model:
— Sundar Pichai (@sundarpichai) October 27, 2026
- 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… pic.twitter.com/BFgQQH9KUb