NewsAdGraam raises ₹3 crore in seed funding
AI Video & Image NewsConfirmed

Genie 4 Is Live in Project Genie for Ultra Users

AdGraam News DeskUpdated 5 min read
A glowing miniature landscape floating above a studio table while a camera rig circles it

The short answer

Google DeepMind said on November 18, 2026 that Genie 4 is now available in Project Genie for eligible Google AI Ultra subscribers, with native audio, real-time text editing and a new Project Genie agent. Google AI Pro users can join a waitlist. Pricing, resolution and an API model ID have not been announced.

On this page (11 sections)

What is Genie 4?

Genie 4 is a world model from Google DeepMind that generates an interactive environment you steer with keyboard controls and text prompts. Google announced it on October 27, 2026, and @GoogleDeepMind said on November 18 that it is now available in Project Genie.

Google calls it its most realistic and interactive world model yet. According to Google, it pairs visuals with real-time native audio that adapts as the scene unfolds, from ambient soundscapes to character dialogue.

A world model differs from a standard video generator. According to DataCamp's write-up, it generates an environment you act inside, frame by frame, instead of rendering a fixed clip.

Genie 4 release date and availability

Genie 4 opened in Project Genie on November 18, 2026, for eligible Google AI Ultra subscribers. Google AI Pro members can join a waitlist, according to a Google developers post quoted by @ai_for_success.

The October 27 announcement described a research preview for trusted testers first. It listed Google AI Ultra subscribers in Project Genie and early access partners of the Gemini API and Google AI Studio as the next channels.

The November 18 post covers the Ultra rollout in Project Genie. Gemini API and AI Studio availability for Genie 4 has not been confirmed as open yet, and Google has not published regions or usage limits.

Genie 4 key features and specs

Genie 4 combines native audio, real-time text editing and Gemini-based reasoning. Gemini Omni anchors real-world appearance and behavior, while other Gemini models keep the simulation coherent. DataCamp names Gemini 3.8 Flash as the guiding agent. Google's pages say only "other Gemini models."

A glass diorama of a foggy lake with a glowing amber sphere and drifting sound wavesA glass diorama of a foggy lake with a glowing amber sphere and drifting sound waves

SpecDetails
MakerGoogle DeepMind
TypeReal-time interactive world model
AudioNative: ambient sound, music, character dialogue
ControlsKeyboard navigation and camera, text prompts mid-session
FoundationGemini family, including Gemini Omni
AccessProject Genie, Google AI Ultra (Pro waitlist)
ResolutionNot announced
Session lengthNot announced for Genie 4
PricingNot announced
API model IDNot announced

The headline feature is mid-session editing. Type a new prompt while you navigate and you can shift the weather, drop in a character or redirect the action. Google says the new Genie 4 audio adapts as the scene changes.

The release also adds a Project Genie agent. You describe a world in conversation, and the agent orchestrates Google's multimodal models to produce initial visuals and a custom background score. Gemini can also watch the screen during a session and suggest where to go next.

Project Genie now includes a community gallery where users publish worlds for others to explore and remix.

What changed for Genie 3

Genie 3 stays in the model picker. Google's listed updates are:

  • Sessions of up to 3 minutes per generation, up from 60 seconds.
  • World chaining, which links multiple worlds into longer journeys.
  • Seeding new worlds from 360-degree panorama images.
  • Real-time in-world events triggered by natural-language prompts.

Google says Genie 4 has the highest-fidelity visuals, physics and audio, while Genie 3 suits longer runs or panorama-seeded spaces.

DataCamp reports that object permanence and legible text remain weak points and that DeepMind says so in its announcement. That matters for any ad that relies on readable packaging.

Genie 4 pricing and free access

Genie 4 has no separate published price. Google ties access to a Google AI Ultra subscription in Project Genie, with a waitlist for Google AI Pro. Google has not announced a free tier or API rates.

DataCamp, writing after the October announcement, said there were no published benchmarks, pricing or API model ID, and that nobody outside Google had evaluated the model. Nothing in the November 18 material changes that.

For models you can run and compare today, the AdGraam Playground lets you run and compare video, image and audio models in one place, and pricing lists credit packs. Genie 4 is not in the Playground.

Genie 4 vs competing AI video models

Genie 4 is made for exploring a live, interactive world. Clip generators produce a video file that you can drop straight into an edit. With Genie 4 you move through a world and change it with keyboard and text input.

Google names training environments for robots and agents, plus filmmaking, journalism, gaming and manufacturing, as target uses. Advertising is not mentioned.

Nothing published shows video file export, vertical 9:16 output or product reference image input. Until Google addresses these, comparing it with clip generators on cost per second, resolution or consistency is not possible.

How to use Genie 4 for AI video and static ads

For ad teams, Genie 4 is worth watching as an idea-testing tool, and it is not ready to produce ads at scale. The plausible fit is exploring a setting, camera path or mood before committing to a shoot or a generated clip.

A creative desk with storyboards and product props beside a monitor showing a dreamlike worldA creative desk with storyboards and product props beside a monitor showing a dreamlike world

Possible early uses:

  • Previsualizing scenes for hooks and brand films.
  • Exploring environments for b-roll concepts.
  • Testing mood with sound, since audio is generated with the scene.

The limits are practical. Weak text legibility rules out label-heavy product shots, and with no export or API documented, there is no way to feed output into a testing budget.

For repeatable ad output now, teams chain models. Flow Builder links image, video and upscale models into reusable workflows, and Video Agent turns a product URL into scripted UGC video ads. Hook testing benchmarks are covered in our hook rate and hold rate guide.

What to watch next

Watch for Gemini API and Google AI Studio access, which would bring a model ID and pricing. Also watch for export formats, resolution and session limits for Genie 4, none of which Google has stated.

The Pro waitlist is the other signal. If it opens, a wider group of creators will start publishing Genie 4 worlds to the gallery, and independent evaluations will follow.

From the timeline

What people are posting

Genie 4 is now available in Project Genie. It brings native audio, real-time editing, and a new agent to help you create interactive, virtual worlds. https://t.co/CipplkOuMP
@GoogleDeepMind293 likesView on X
Google is just warming up! 🔥 Google has now started rolling out Project Genie with Genie 4 to Google AI Ultra subscribers. Genie 4 includes real-time world editing with native audio. https://t.co/U150EjpXTO
@ai_for_success1 likeView on X
GOOGLE 🔥: Genie 4 is now rolling out on Project Genie for Google AI Ultra subscribers. > Genie 4 comes with native audio and realtime editing functionalities. > Users can also explore a gallery of pre-made creations to explore and remix. Ultra soon 👀 https://t.co/nrxrcBKNsK
@testingcatalog12 likesView on X

Key takeaways

  • Google DeepMind announced Genie 4 on October 27, 2026 and opened it in Project Genie on November 18, 2026.
  • The rollout covers eligible Google AI Ultra subscribers, with a waitlist for Google AI Pro.
  • Genie 4 generates ambient sound, music and character dialogue natively alongside the world.
  • Users can type prompts mid-session to change weather, add characters or redirect the action.
  • Genie 3 stays in the model picker with sessions of up to 3 minutes, up from 60 seconds.
  • No pricing, resolution, export format or API model ID has been published.
Run today's top models

Compare 200+ AI video and image models side by side

Run Kling, Veo, Sora, Seedance, Flux, Ideogram and Recraft in one Playground, or turn any product URL into ready-to-launch ads.

Questions, answered

Frequently asked

When did Genie 4 come out?

Google DeepMind announced Genie 4 on October 27, 2026 as a research preview for trusted testers. On November 18, 2026 it said Genie 4 is now available in Project Genie for eligible Google AI Ultra subscribers.

How do I get access to Genie 4?

Eligible Google AI Ultra subscribers can use it through the Project Genie website. Google AI Pro members can join a waitlist. Early access partners of the Gemini API and Google AI Studio are listed as a later channel.

How much does Genie 4 cost?

Google has not published a separate price for Genie 4. Access is tied to a Google AI Ultra subscription in Project Genie. API pricing has not been announced.

Is Genie 4 free to try?

Not as far as Google has said. The full experience is rolling out to Ultra subscribers, and Pro users are on a waitlist. No free tier has been announced.

Can you use Genie 4 for commercial ads?

Google has not published commercial terms or an export format for Genie 4. Its announcement names filmmaking, journalism, gaming and manufacturing as creative uses, and does not mention advertising.

How is Genie 4 different from Genie 3?

Google calls Genie 4 its most realistic and interactive world model, and the first Genie with native real-time audio. Genie 3 remains selectable for longer runs of up to 3 minutes and for 360-degree panorama seeding.

Where this comes from

Sources

  1. Project Genie update featuring Genie 4blog.google
  2. Genie 4: A new generation of world simulationdeepmind.google
  3. Genie 4: Google DeepMind's New Real-Time World Modeldatacamp.com
  4. We're announcing Genie 4, our new world model for real-time interactive ...blog.google
AI Video & Image Newsgenie 4project geniegoogle deepmind world modelgenie 4 release dategenie 4 accessgoogle ai ultraai world modelgenie 3 vs genie 4
Keep reading

More AI model news & creative guides

AI Video & Image News

SenseNova U1 Pro: No. 6 on DesignArena, Native 8K

SenseNova U1 Pro from SenseTime debuts No. 6 in DesignArena image generation and editing, with claimed native 8K output. Get specs, ad workflow notes and access.

Nov 19, 2026 · 5 min
AI Video & Image News

Recraft Studio Font in Prompt: Attach Any Typeface

Recraft Studio now lets you attach any typeface, from its Font library or your own brand font, to a prompt. Get the ad workflow breakdown and access details.

Nov 18, 2026 · 4 min
AI Video & Image News

Muse Code 3D: Native GLB Generation, Limits & Access

Meta added native text and image to GLB 3D generation to Muse Code on Nov 20, 2026. Get the spec table, ad workflow breakdown, plan limits and access details.

Nov 21, 2026 · 5 min
AI Ad Creation

Best AI Ad Generators for D2C Brands in 2026

Ten AI ad generators compared for D2C and e-commerce brands: what each makes, real starting prices, free options, and which suits Indian brands paying in rupees.

Nov 7, 2026 · 11 min
Meta Ads for Indian D2C

Break Even ROAS for COD Orders: How RTO Changes the Math

How to work out break even ROAS when a share of COD orders comes back as RTO, with worked rupee examples, a prepaid and COD mix, and why Ads Manager ROAS overstates COD.

Nov 7, 2026 · 7 min
Meta Ads for Indian D2C

What Indian D2C Brands Run on Meta: 2,790 Ads Analyzed

We studied 2,790 Facebook and Instagram ads from 14 Indian brands and their 747 longest-running ads: formats, angles, offers, languages and CTAs that last.

Nov 7, 2026 · 7 min

Your move

See your product as a winning ad

Turn a product URL into static ads and UGC-style videos, rebuild the competitor ads that already work, and read your Meta data to know what to make next.

AdGraam News DeskAI video and image news

The AdGraam News Desk (part of AdGraam, also searched as Adgram or Adgram AI) tracks AI image and video model launches, leaks and rumors from labs, leakers and creators, and checks each claim against its sources. Unconfirmed stories say so, and each page is updated as its story develops.

LinkedIn