Google’s premium Gemini users can now generate fully playable AI worlds with Project Genie

What happened: Google has rolled out Project Genie for its Gemini Ultra subscribers in the US, enabling them to create playable AI-generated virtual worlds.

Why it matters now: This marks the first consumer-facing release of Genie 3, Google DeepMind’s most advanced world-generation model.

What changes for people: Premium users can now build entire 2D and 3D interactive environments using simple prompts or images.

Who is affected: Gemini Ultra subscribers, AI creators, developers, virtual world designers, and users exploring next-gen AGI tools.

Google unlocks AI-generated playable worlds for premium Gemini users

In a major step toward next-gen immersive AI experiences, Google has announced that Project Genie is now available for Gemini AI Ultra subscribers in the US.

The feature allows users to instantly generate fully interactive, physics-based virtual worlds simply by typing prompts or uploading images.

Project Genie is powered by Genie 3, a world-generation model developed by Google DeepMind, along with support from Nano Banana and Gemini 3.

This is one of the first times Google has allowed the public to access world-building models originally designed to support its broader AGI mission.

The Ultra subscription is priced at $249.99 per month in the US. In India, the plan is listed at ₹24,500, but Project Genie is not yet available for Indian users.

What Project Genie can do

Project Genie brings a set of powerful creation tools that allow users to:

generate 2D or 3D virtual worlds

explore environments using walking, flying, riding, or driving modes

create custom characters with their own movement styles

interact with real-time rendered environments

experience realistic physics, even during dynamic gameplay

The system can simulate terrain, buildings, creatures and objects, while also predicting movement ahead in real time as the player navigates the world.

World Sketching: Create environments with fine control

The feature also includes World Sketching, powered by Nano Banana Pro, offering granular environment editing. Users can:

preview worlds before entering

tweak layout, lighting, textures and shapes

design landscapes using both sketches and prompts

switch between first-person and third-person views

This allows creators to craft detailed, customised, game-like environments without any coding or game-engine expertise.

Remix mode: Build on your earlier creations

Another highlight is the Remix tool, which lets users revisit their previously generated worlds and:

expand them

add new elements

change terrain

experiment with alternate styles

evolve storylines or character paths

This makes Project Genie not just a tool for one-off creations but an evolving sandbox for interactive AI experimentation.

Why this matters for AI and gaming

Genie 3 is one of the first models capable of combining:

world generation

real-time physics

realistic interaction

playable navigation

both 2D and 3D rendering

Google says these capabilities will influence:

robotics simulation

animation and motion modelling

game development pipelines

interactive storytelling

AGI research requiring dynamic environments

For creators, it offers a completely new way to prototype virtual worlds without game engines like Unity or Unreal.