
Written By Deepti Ratnam
Published By: Deepti Ratnam | Published: Aug 06, 2025, 12:23 PM (IST)
Google’s DeepMind is at the centre stage of evolving artificial intelligence breakthroughs. The tech giant is not just building advanced models that push boundaries of artificial intelligence, but also keeping it at the breakneck speed of what machines can create and understand. In a latest update, the tech giant has launched Genie 3, a real-time AI-generated model that transforms and evolves the gaming and virtual experience.
Genie 3 is a world model AI capable of generating interactive 3D environments in real time, based entirely on text prompts. It helps in creating new ways of immersive and dynamic videos that feel like bringing science fiction to reality.
The model is made to offer a more immersive and consistent experience compared to previous versions from Google. It works closely with retaining environmental details for longer and enabling continuous interaction.
The new model, Genie 3, runs at 24 frames per second along with a resolution of 720p, hence keeping the visuals smooth and responsive for a longer time.
Additionally, Genie-3 is also using auto-regressive frame generation that takes reference from past frames in order to provide consistency. This ensures that if users revisit a location, the environment remains accurate, preserving context for up to a minute.
Currently, Google has not announced any dates for the launch of the Genie-3 for the public. However, the official blog post says, the company is ‘exploring how we can make Genie 3 available to additional testers in the future.’