Skip to main content
MindStudio
Pricing
BlogAbout
My Workspace
Project Genie priceGoogle AI Ultra costProject Genie access region

Project Genie Pricing: What It Costs to Access Google's World Model

Project Genie requires a Google AI Ultra subscription near $100/month plus regional access via labs.google/fx. Here's what that gets you.

Edited by Luis Chavez-Mattos, Director of Product RSS
Project Genie Pricing: What It Costs to Access Google's World Model

What does Project Genie cost right now?

Project Genie sits behind Google’s Ultra subscription tier, priced at roughly $100 a month, and it’s only reachable through a specific experimental URL (labs.google/fx/projectgenie) rather than any mainstream Google product page. On top of the subscription cost, access is currently limited to certain regions, so paying for Ultra doesn’t guarantee entry if you’re outside the supported countries. There’s no separate, cheaper tier for Genie alone. If you want in, Ultra is the only door.

TL;DR

  • Project Genie requires the Google AI Ultra plan, which starts at around $100 per month, and there’s no standalone or discounted way to access it separately.
  • Regional restrictions apply on top of the subscription, meaning even Ultra subscribers outside supported countries (the US being one confirmed example) can’t currently get in.
  • The access point is a Google Labs URL, specifically labs.google/fx/projectgenie, underscoring that this is an experimental release, not a polished consumer product.
  • Genie generates interactive 3D environments in real time, not pre-rendered video, so areas outside your immediate view literally don’t exist until the model generates them.
  • The experience is rough by design, with choppy movement and visual glitches that make it look more like an early tech demo than a finished game engine.
  • Several sample simulations are included, ranging from a dinosaur safari to a social scenario generator, giving early users a sense of possible use cases beyond entertainment.
  • Nobody has fully figured out the killer use case yet, though early speculation points toward training simulations, social skills practice, and experiential education.
Developers have their tools. Enterprises have Remy.
Developers · AI coding tools
Claude Code
Codex
Cursor
GitHub Copilot
Enterprises · built into Remy
⌂Self-hosted
✓SOC 2
✚HIPAA
☰Audit log

What exactly is Project Genie?

Project Genie is Google’s generative world model, a system that creates interactive 3D environments from text prompts and then keeps generating new parts of that environment as a user moves through it. This is a different category from AI video generation tools like Sora or Veo. Those produce a finished clip. Genie instead produces a persistent, explorable space where the content you haven’t looked at yet doesn’t exist until the model renders it on the fly.

That live-generation approach means the experience can look rough. Movement is choppy, objects can glitch or transform unexpectedly, and the whole thing resembles an early-stage video game more than a commercial release. That’s not an accident of bad engineering so much as the nature of generating an entire 3D world frame by frame in response to user input. It’s an early-stage research preview, not a shipped product, and it looks like one.

Why is access locked behind Ultra and region restrictions?

Google has a pattern of gating its most experimental or compute-heavy tools behind its highest subscription tier and a limited rollout footprint. Project Genie fits that pattern on both counts. It lives under Google Labs, the division Google uses to test early-stage ideas before (or instead of) folding them into mainstream products like Search, Workspace, or Gemini. Labs releases are explicitly framed as experiments, and Genie’s placement there signals Google isn’t ready to commit it to a general audience yet.

Generating a 3D world in real time is almost certainly expensive from a compute standpoint, more so than generating a static image or even a short video clip, since the model has to continuously produce new environment data as a user interacts with it. Pairing that cost with the Ultra tier, and limiting it geographically, lets Google control how many people are hammering the system at once while it’s still unstable.

How do you actually get access?

To try Project Genie, you need two things lined up at once: an active Google AI Ultra subscription, and your account needs to be in a region where the feature has been enabled. The US was confirmed as one supported region. From there, the access point isn’t inside the main Gemini app or Google’s consumer AI hub. It’s a separate experimental page at labs.google/fx/projectgenie.

That URL structure matters. Anything under labs.google/fx has historically been Google’s sandbox for testing generative tools before deciding whether they graduate into broader products. Treat the URL itself as a signal: this is pre-release, likely to change shape, and not guaranteed to stick around in its current form.

Is Project Genie worth paying for?

Build enterprise applications with AI that are
Secure.Role: Adjuster✕ Export PHI
Regulation-compliant.✓ HIPAA✓ SOC 2
Auditable.09:14M. Chen approved claim #4471
Scalable.12,480 users online

That depends entirely on why you’d pay. If you’re hoping for a polished, game-like experience, the answer is no. The movement is awkward, the generated geometry can be inconsistent, and there’s no clear “product” here yet in the way there is with, say, an image generator or a coding assistant. If you’re subscribing to Ultra purely to try Genie for a day and then cancel, you’re paying $100 for a tech demo.

But if you’re already evaluating frontier AI tools for other reasons, or you’re the kind of person who wants to see new capabilities before they’re refined, Genie is one of the more genuinely novel things Google has shipped recently. It’s not another chatbot update or incremental model bump. It’s a different category of generative output: a persistent, navigable space instead of a static artifact. Early, rough versions of transformative tools have historically looked unimpressive (early text generation models were far worse than what’s available now), so the rough edges aren’t necessarily evidence the underlying idea is a dead end.

What might Project Genie actually be used for?

Nobody, including Google, has settled on a definitive answer yet. The demos available through the Labs access point hint at a few directions: simulated environments for exploration (a dinosaur safari, a walkable construction site), social or conversational scenarios where AI-voiced characters interact with the user, and lighter entertainment formats like an endlessly scrolling generated video feed.

The more interesting speculation, floated by early testers rather than confirmed by Google, is around training and education. A world model that can simulate physical environments and generate characters that respond naturally could plausibly be used to rehearse real-world scenarios, practicing a difficult conversation, running through a business decision, or simulating a physical environment before encountering it for real, rather than just reading about it. None of that is a shipped feature today. It’s a reasonable extrapolation of what a generative world model makes technically possible.

Frequently Asked Questions

How much does Project Genie cost?

Access requires a Google AI Ultra subscription starting at roughly $100 per month. There is no separate or cheaper plan specifically for Genie.

Can anyone with Ultra access Project Genie?

No. Access also depends on region. Ultra subscribers outside supported regions currently cannot reach the tool even with an active subscription.

Where do I go to try Project Genie?

The access point is labs.google/fx/projectgenie, a page under Google’s experimental Labs division rather than inside the main Gemini consumer app.

Is Project Genie the same as Google’s video generation tools?

No. Video generators like Veo produce finished clips. Genie generates an interactive, explorable 3D environment that builds itself in real time as the user moves through it.

Is Project Genie a finished product?

No. It’s explicitly an early-stage experiment hosted under Google Labs, with rough visuals, inconsistent object behavior, and no confirmed roadmap for a wider or permanent release.

Editorial standards

Presented by MindStudio

No spam. Unsubscribe anytime.