Google Genie 4 Builds Worlds You Can Steer Live
Google DeepMind has unveiled Genie 4, its most realistic world model yet, which builds interactive worlds with real-time audio that users can steer by typing.
Google DeepMind has released Genie 4, its newest world model, as a research preview. Unlike a video generator that plays a fixed clip, a world model builds a scene you can move through in real time. Genie 4 goes further than its predecessor: it creates more photorealistic worlds, adds sound that matches the scene, and lets you change the scene while it is running, simply by typing.
What Genie 4 can do
Genie 4 generates interactive worlds from text or image prompts. Users move through a scene with keyboard controls and steer it with plain-language prompts typed mid-simulation. You can change the weather, move the camera to a different angle, or add new people or objects, without restarting the scene. The model also generates real-time audio to match what is happening on screen. That includes ambient sounds, music, and dialogue from characters inside the world. Google describes Genie 4 as its most photorealistic and interactive world model yet, with better physics and more human-like characters than Genie 3.
Longer, steadier worlds
One of the clearest upgrades is staying power. Genie 4 can keep a scene consistent for up to two minutes, double the limit of Genie 3, which Google released as a research preview a little over a year ago. That matters because a world model has to remember what is where: if you turn around and turn back, the world should still look like the same place. Google's Gemini model is also more closely integrated with Genie 4. Gemini acts as a director, planning steps and checking frames while Genie creates the visuals. In a research experiment tied to Gemini Robotics, a Gemini model broke tasks into steps that Genie 4 then simulated in real time, showing how robot scenarios could be tested in simulation before they happen in the real world.
Who can use it, and when
For now, Genie 4 is available only as a research preview for trusted testers. Google says it will roll out to Google AI Ultra subscribers through Project Genie in the coming weeks, and to early-access partners through the Gemini API and Google AI Studio. Google is pitching the model for gaming, interactive learning, and robotics. Trusted testers are also pointing to a newer use: prototyping films, where a director can step into a live scene and try ideas quickly during storyboarding and early ideation.
Still early days
Google is clear that Genie 4 is not a finished product. Engineers say it is still early days and the final product experience is not yet clear. Details and objects can drift the longer a simulation runs, and the team says accurate audio and legible text inside the world are still problems to solve. The release also lands in a busy field. Runway released its second world model, GWM Worlds 2, a couple of months ago, and the wider world-model industry is consolidating, with acquisitions and talks involving World Labs, Decart, and Third Dimension. Genie 4 is Google's latest move to stay ahead in that race, but for most people, the wait for hands-on access continues.
Sources
- TechCrunch - https://techcrunch.com/2026/10/27/googles-genie-4-world-model-adds-audio-more-realistic-worlds-and-real-time-steering/ - Times Now - https://www.timesnownews.com/technology-science/google-unveils-genie-4-ai-model-that-creates-realistic-worlds-with-sound-and-music-article-156419372