genie 4
Type a sentence, get a world you can walk around: Genie 4 renders the scene, simulates the physics and voices every character in real time.
Feedback
AI Dance Tools
Explore our AI-powered dance video creation tools
MiniMax H3
MiniMax H3 AI Video Generator
Seedance 2.5
The Future of AI Video Is Here.

Seedance 2.0
The Future of AI Video Is Here.

Veo3.1
Create Stunning Videos with Veo3.1
FLUX 3 Video Generator

Kling 3.0
Next-Gen AI Video Generator
Grok Video Generator
Create Videos from Text or Images with AI
MiniMax H3 video generator
MiniMax H3 AI Video Generator
The Simulation Engine Behind Genie 4
Hand the model a written brief or a starting frame, and it answers with a scene that keeps running — picture, movement, sound and speech all from the same pass.
- Everything Generated in One PassPicture, ambience and spoken lines are produced together from the start, so audio never feels bolted on after the Genie 4 scene is finished.
- Steer With Plain LanguageCall for rain, a wider lens or a second character mid-session — Gemini applies the change to whatever is already on screen.
- Powered by the Gemini LineupGemini Omni anchors the visuals and behavior in physical reality, while companion Gemini models track continuity between frames.
Four Steps to Your First Genie 4 World
Waitlist signup to a scene you can steer, in four moves.
Capabilities Packed Into Genie 4
Six things this generation of real-time world simulation gets right.
Weight, Momentum and Collisions
Objects carry convincing mass — think airflow, hard impacts and the small physical cues that make a scene read as real.
Human Motion That Reads True
Figures walk, gesture and speak like people do, and you can nudge their behavior while the simulation keeps running.
Sound Baked Into the Scene
Every location ships with its own soundscape, from wind through branches to waves hitting the shore, generated with the frame.
Characters That Answer You
Speak or type and the cast responds in the moment, with lines that shift as circumstances inside the world change.
Planning Across Many Steps
Gemini 3.8 Flash sits in the loop, mapping what happens next, checking each frame and rerouting when events take a turn.
A Camera That Looks Photographic
Light, grain and lens behavior stay consistent across long takes and free camera moves, so footage holds up on a big screen.
Genie 4: Questions Answered
Straight answers on access, audio, controls and what Genie 4 is built for.
So what is Genie 4, exactly?
A multimodal world model — given a prompt or a frame, it predicts how that scene should look, sound and behave next.
How can I try it?
Through the Project Genie waitlist for the research preview, or early access via the Gemini API and Google AI Studio.
Will I get audio and talking characters?
You will — ambience and dialogue are produced by the model itself and shift in response to what you do.
What controls do I have?
Keyboard input moves you around; typed prompts redirect weather, camera or cast while the scene plays on.
Where do the Gemini models fit in?
Gemini Omni keeps appearance and motion believable; the other Gemini models maintain consistency as time passes.
What can people build with it?
Robot and agent training grounds, navigation benchmarks for AI, and interactive experiences that were not possible before.
Get Early Access to Genie 4
Sign up for the Project Genie waitlist, or reach the model sooner through Google AI Studio and the Gemini API.
