Google Genie 4 launches in Project Genie with live edits and real-time sound
By The Daily Byte
·20 Nov 2026
·4 min read

Google has opened Genie 4, its latest world model, to eligible Google AI Ultra subscribers through Project Genie, its browser-based prototype for generating and exploring interactive worlds. The rollout, announced on 18 November 2026, is the first time Genie 4 has been available beyond a small group of trusted testers.
Genie 4 lets users change a generated world while they are still moving through it. Typing a new prompt mid-session can shift the weather, drop in a character or redirect the action, without starting the world again. It is also the first model in the Genie line to pair its visuals with real-time native audio that adapts as the scene unfolds, from ambient soundscapes to character dialogue.
Google describes Genie 4 as its most realistic and interactive world model yet. The update also brings a new Project Genie agent and a community gallery, aimed at making the prototype easier to build with and to share.
## What Genie 4 changes inside a world
Earlier Genie models generated explorable scenes from a text or image prompt, but the world was largely fixed once it began. Genie 4 adds live control: users can type prompts while navigating to edit environments in real time. Google says the model combines its world-model research with Gemini models, which supply reasoning to keep scenes coherent as they change.
The audio is the other step forward. Where Genie 3 worlds were silent, Genie 4 generates sound in the same pass as the visuals, so music, background noise and spoken dialogue react to what the user does and what happens in the simulation.
Google is careful to frame both models as frontier research systems. It says it is still working on camera latency, intricate physical interactions, real-time audio precision and long-term scene permanence, and that details in a world can still shift or drift during longer sessions.
## Genie 3 gets longer sessions and panorama worlds
Genie 3 remains available in the Project Genie model picker, and Google says it is the better choice for longer runs or worlds seeded from 360-degree panorama images. Its sessions have been extended to up to three minutes per generation, up from 60 seconds, and multiple worlds can now be linked together for longer journeys.
Genie 3 has also gained 360-degree panorama image grounding, so a panorama photograph can seed a new world, and real-time in-world events that change the environment through natural-language prompts while the user explores.
## A conversational agent and a shared gallery
The new Project Genie agent takes creation beyond a single prompt. Users describe their vision in natural language and the agent orchestrates Google's suite of multimodal models to build the opening visuals and a custom background score matched to the mood.
Gemini models can also join a session as the user explores, observing what is happening on screen, surfacing context-aware suggestion chips for where to go next, or triggering in-world events automatically.
The community gallery brings together curated worlds and creations from the community. Google highlights "Living Memory", which turns personal photographs into explorable spaces, "Dino Field Guide", a prehistoric habitat grounded in 360-degree panoramas, and "Rock Paper Scissors", in which Gemini acts as a real-time referee. Early collaborations include "Villa Giallo", a gothic manor mystery built with creative technologist Will, known as BLVCKL!GHT, and "The Life of Ants", built with filmmaker Anand Gandhi, which invites users to navigate an ant colony from an ant's perspective.
## Who can try it, and the safety measures
Eligible Google AI Ultra subscribers can try the new Project Genie experience from 18 November 2026, while Google AI Pro members can join a waitlist on the same page. Google has not said when Genie 4 will reach other subscription tiers or regions beyond where Google AI Ultra is available.
All video and audio generated in Project Genie carries SynthID, Google's imperceptible watermark for identifying AI-generated content, and all model inputs and outputs go through safety screenings. Google DeepMind first announced Genie 4 as a research preview for trusted testers on 27 October 2026; the 18 November rollout is the next step in that plan, bringing the model to paying subscribers in Project Genie.
## Frequently asked questions
**What is Google Genie 4?**
Genie 4 is Google DeepMind's latest world model. It generates interactive, explorable worlds in real time and is the first Genie model to add native audio and mid-session editing.
**How is Genie 4 different from Genie 3?**
Genie 4 offers higher-fidelity visuals and physics, real-time native audio and live prompt editing. Genie 3 remains available for longer runs and panorama-seeded worlds.
**Who can use Genie 4 now?**
Eligible Google AI Ultra subscribers can use it in Project Genie from 18 November 2026. Google AI Pro members can join a waitlist.
Sources: [Google — Genie 4 is now available](https://blog.google/innovation-and-ai/models-and-research/google-deepmind/genie-4-now-available-project-genie/) and [Unite.AI — Google Brings Genie 4 to Project Genie for Eligible AI Ultra Users](https://www.unite.ai/google-brings-genie-4-to-project-genie-for-eligible-ai-ultra-users/).
Source: Google Blog.