# Google releases Genie 4 for AI Ultra subscribers alongside new world-building tools

> Google is rolling out Genie 4 to AI Ultra subscribers, letting them edit 3D worlds in real time with natural language prompts.

- Type: AI news
- Page: https://completeaitraining.com/news/google-releases-genie-4-for-ai-ultra-subscribers-alongside/
- Published: 2026-11-19
- Updated: 2026-11-19
- Author: Jeroen Erne
- Jobs: Creatives, Hospitality and events, Marketing, Product development, Writers
- Reading time: 2 min

Google has begun rolling out Genie 4 to Google AI Ultra subscribers, upgrading its Project Genie platform with a model that lets users edit interactive 3D worlds in real time using natural language prompts. The update, announced November 18, 2026, also adds native audio generation and a community gallery for sharing and remixing creations.

Project Genie first launched earlier this year as an experimental research prototype. Early testers asked for more control over their generated environments and simpler ways to build and share them. "We are rolling out Genie 4 to eligible Google AI Ultra subscribers, offering more responsive, social, and interactive world generation," said Diego Rivas, Group Product Manager at Google DeepMind, and Pavan Kumar, Group Product Manager at Google Labs.

#### What Genie 4 brings to the table

Genie 4 is the most realistic world model Google has released. Users can type prompts mid-session to shift weather, drop characters into a scene, or redirect the action while navigating their world. For the first time, these visuals are paired with adaptive native audio - ambient soundscapes and character dialogue that change as the scene unfolds.

The update also introduces the Project Genie agent, which lets users build complex environments through conversation. Describe your vision in natural language, and the agent coordinates Google's multimodal models to establish a world's visual signature and pair it with a custom background score. Gemini models can also observe the action on screen, surfacing suggestion chips for where to go next or triggering in-world events automatically.

#### Genie 3 upgrades and new creative tools

Genie 3 remains available in the model picker and received several upgrades alongside the Genie 4 launch. Sessions now run up to three minutes, up from 60 seconds, and users can chain multiple worlds together for longer journeys. A new 360-degree panorama image grounding feature lets you seed worlds from panoramic photos, keeping environments like Petra's canyon consistent as you look around. Real-time promptable events allow dynamic environment changes while exploring.

The new community gallery offers curated worlds and community creations. Examples include "Living Memory," which transforms personal photos into explorable spaces, and "Dino Field Guide," built from 360-degree panoramas of prehistoric habitats. Google also collaborated with creative technologist Will, known as BLVCKL!GHT, on "Villa Giallo," a room-by-room mystery set in a gothic manor, and with filmmaker Anand Gandhi on "The Life of Ants," which lets users navigate a colony from an ant's perspective using Genie 4's photorealism.

#### Safety and research limitations

All video and audio generated in Project Genie includes SynthID watermarking, and model inputs and outputs go through safety screenings. Google acknowledges that Genie 3 and Genie 4 remain frontier research models. The team is working on challenges including camera latency, intricate physical interactions, real-time audio precision, and long-term scene permanence. A full research overview details these limitations.

Google AI Pro members can join a waitlist, while Ultra subscribers can start building immediately on the Project Genie website. For creative professionals looking to build skills with generative tools, [AI for Creatives Courses](https://completeaitraining.com/tag/creatives/) and [Generative AI Courses](https://completeaitraining.com/tag/generative-ai-and-llm/) offer practical training on the underlying technologies shaping these platforms.

#### Why this matters for creatives, marketers, and product developers

Genie 4 shifts world generation from a one-shot output to an iterative, conversational process. A marketing team can prototype an immersive brand experience, adjust the mood and setting in real time during a client call, and publish a remixable version for community engagement - all without touching a 3D modeling tool. Writers and event designers can use the agent to block out spatial narratives, testing how a scene feels with different lighting, weather, and audio before committing resources to production. The gallery's remix feature means you can fork someone else's world, tweak the rules or soundtrack, and publish your own variant, which opens a low-friction path to collaborative prototyping across teams.

---

This is the Markdown form of https://completeaitraining.com/news/google-releases-genie-4-for-ai-ultra-subscribers-alongside/. The whole site is mapped at https://completeaitraining.com/llms.txt.
