🇺🇸 English 🇰🇷 한국어 🇯🇵 日本語
Home › AI & Machine Learning
AI & Machine Learning

Genie 4 Opens to Google AI Ultra Subscribers

Google on Nov. 18 opened its Genie 4 world model to eligible Google AI Ultra subscribers through a redesigned Project Genie, adding real-time world editing, native audio, a creation agent and a community gallery. Developer API access has no date yet.

Google on Nov. 18 opened Genie 4, its newest world model, to eligible Google AI Ultra subscribers through a redesigned Project Genie, ending a three-week stretch in which only invited testers could use it. The update also adds a conversational creation agent, a community gallery and longer sessions for the older Genie 3 model.

Aerial view of Google's main campus buildings in Mountain View, California
File photo: Google's main campus in Mountain View, California. 사진: Austin McKinley / 위키미디어 공용(CC BY 3.0)

Key takeaways

  • Genie 4 is now available to eligible Google AI Ultra subscribers in Project Genie; Google AI Pro members can join a waitlist.
  • Users can type prompts mid-session to change weather, add characters or redirect action, with native audio that adapts in real time.
  • Genie 3 stays in the model picker with sessions of up to 3 minutes per generation, up from 60 seconds.

What Genie 4 in Project Genie can do

In a post on The Keyword, Google DeepMind group product manager Diego Rivas and Google Labs group product manager Pavan Kumar called Genie 4 "our most realistic and interactive world model yet." A world model generates an explorable environment frame by frame and reacts to what the user does, rather than playing back a fixed clip.

Inside Project Genie, users can type a new prompt while moving through a world to "shift the weather, drop in a character, or redirect the action." Google said Genie 4 pairs those visuals with real-time native audio, "from ambient soundscapes to character dialogue."

New Project Genie agent and gallery

The update adds a Project Genie agent. Users describe a world in plain language, and the agent coordinates Google's multimodal models to produce opening visuals and a custom background score. Gemini models can also join a session, watching the screen, offering suggestion chips and triggering in-world events.

A new gallery mixes curated and community-built worlds. Examples include "Living Memory," which turns personal photos into explorable spaces, and "The Life of Ants," built with filmmaker Anand Gandhi. Users can publish their own worlds for others to explore and remix.

Genie 3 upgrades: longer sessions, panoramas

Genie 3 remains available for longer runs. Google raised its limit to 3 minutes per generation from 60 seconds, added "world chaining" to link several worlds, and now lets users seed worlds from 360-degree panorama images. Real-time promptable events, which were missing at launch, are now supported.

Context: from research preview to subscription

Google DeepMind announced Genie 4 on Oct. 27 as a research preview for trusted testers, with Gemini 3.8 Flash acting as a guiding agent that plans and checks each simulation. It said Ultra subscribers and early-access partners of the Gemini API and Google AI Studio would follow. AI Newsway covered that announcement on Oct. 28.

Project Genie itself launched on Jan. 29 for U.S. Ultra subscribers aged 18 and older, running on Genie 3 with generations capped at 60 seconds. Google DeepMind first previewed Genie 3 in August 2025.

All video and audio generated in Project Genie carries Google's SynthID watermark, and model inputs and outputs go through safety screening, the company said.

What's next

Google's Nov. 18 post does not mention developer access, and it gives no date for the Gemini API and AI Studio rollout DeepMind described in October. Google lists camera latency, complex physical interactions, real-time audio precision and long-term scene permanence as open challenges.

FAQ

Who can use Genie 4 now?

Eligible Google AI Ultra subscribers through the Project Genie website. Google AI Pro members can sign up for a waitlist on the same page.

Can developers use Genie 4 through the Gemini API?

Not yet announced. DeepMind said in October that early-access API partners would follow, but Google has not given a date.

How long can a session last?

Google gave no Genie 4 session length. Genie 3 sessions now run up to 3 minutes per generation.

References

#Genie 4 #Project Genie #Google DeepMind #world model #Google AI Ultra #Genie 3 #SynthID #interactive world generation
S
Seung Jung
AI Newsway · Developer

Editor responsible for AI Newsway. Writes about AI, developer tools, cloud and productivity technology.

More by this reporter ›
Copyright © AI Newsway All rights reserved. Unauthorized reproduction, redistribution, and AI training prohibited.
Share
f X K

Comments 0

Comments are currently disabled.

Be the first to comment.