Genie 4 Steps Out of the Lab and Into Paying Hands

Human Verified

Share to: , and

A person walking through a luminous computer-generated valley of floating islands and waterfalls at sunset

Google opened Genie 4 to eligible AI Ultra subscribers on November 18, moving its most realistic world model out of a trusted-tester preview and into the redesigned Project Genie. The rollout adds a creation agent and a community gallery, plus longer Genie 3 sessions.

First Access for Ultra Subscribers

The new Project Genie experience is available to eligible Google AI Ultra subscribers starting today, Google said. Google AI Pro members can join a waitlist on the same page, while developers still wait for word on Gemini API or AI Studio access.

Genie 4 was announced on October 27 as a research preview for trusted testers, so the Ultra rollout is the first time paying users can walk through its worlds. Google introduced the model as its most realistic and interactive world model yet.

What Genie 4 Adds

With Genie 4, users can type a new prompt mid-session to change the world they are exploring, shifting the weather, adding a character or redirecting the action in real time. It is the first Genie model with native audio that adapts as the scene unfolds.

The audio ranges from ambient soundscapes to character dialogue, Google said, and the model keeps the world coherent over short sessions. Details can still drift in longer runs or after several complex edits, a limit Google openly acknowledges in its research overview.

Genie 3 Gets Longer Sessions

Genie 3 remains available in the model picker for longer runs or panorama-seeded spaces, and its sessions now stretch to three minutes per generation, up from 60 seconds. Worlds can also be chained together for longer journeys across multiple connected scenes.

New Genie 3 features include 360-degree panorama image grounding, which seeds worlds from panorama photos, and real-time in-world events that change the environment through natural-language prompts while users explore. The upgrade keeps the older model useful beside its flashier successor.

The new Project Genie agent turns creation into a conversation: describe a vision in natural language and it orchestrates Google’s multimodal models to set the look and a custom background score. Gemini models can also watch the screen and suggest where to go next.

The community gallery gathers curated worlds and community creations to explore and remix, from turning personal photos into explorable spaces to a prehistoric habitat tour and a desk game with Gemini as referee. Creators can publish their worlds for others to remix.

Our Read

Genie 4 reaching paying users in 22 days is the fastest productization in the Genie series, and it shows Google treating world models as a consumer product, not just a research demo. The agent and gallery matter as much as the model itself.

The honest caveats are doing real work here: drift, latency and permanence limits are stated up front, and SynthID watermarking covers all generated video and audio. The open question is whether anyone builds lasting experiences in a world that still shifts underfoot.

Why It Matters

World models are becoming the next front in generative AI, with AMD buying World Labs and rivals racing to make explorable spaces. Google putting Genie 4 in subscribers’ hands first, with sound and live edits, sets the pace others must now match.

Background

Project Genie began in January 2026 with Genie 3 for AI Ultra subscribers in the United States, and Genie 4 arrived as a research preview on October 27. Gemini 4 Neon launched the same month, keeping DeepMind’s releases on a fast cadence.

Source: Google’s Genie 4 update

Share to: , and

Faheem R. Khattak

Senior automation engineer and AI marketing journo covering LLMs, algorithms, analytics, core updates, Bing search, and technical SEO.

0 Comments

Leave a Reply

Optional Take Tag:

🛡️ Spam protection: Honeypot + Time-Trap Active

This site uses Akismet to reduce spam. Learn how your comment data is processed.