On November 18, 2026 (November 19, 2026 at 02:15 Beijing time), Google announced via its official X account (@Google) that Genie 4, Google DeepMind's interactive world model, is now available in Project Genie. Following its initial research preview and waitlist debut on October 27, 2026, Genie 4 arrives in Project Genie alongside new co-creation tools, a community gallery, and upgrades to Genie 3.

Four Core Updates in Project Genie

Genie 4 in Project Genie: **Users can edit simulated environments mid-session using text prompts, paired with real-time native audio and dialogue. Per Google DeepMind's model page, Genie 4 builds on the Gemini family—using **Gemini Omni to anchor simulations in real-world appearance and behavior while Gemini reasoning models maintain coherence and continuity.

**Project Genie Agent: **Enables conversational world co-creation with custom soundtracks and live Gemini guidance.

**Community Gallery: **Lets users discover, explore, and remix interactive worlds shared by creators and the Project Genie community.

Genie 3 Upgrades: **Extends interactive sessions up to **3 minutes and introduces world chaining alongside 360-degree panorama grounding.

Figure 1 Overview of the four core feature modules in Project Genie with Genie 4
Figure 1 Overview of the four core feature modules in Project Genie with Genie 4

Image source: Conceptual overview diagram created by DigIntoAI based on the official Google announcement and Google DeepMind model page.

Interactive Demo and Availability Scope

In a 58-second demonstration shared with the announcement, a user navigates an interactive scientific laboratory scene using WASD and Spacebar controls to observe and interact with a robotic arm examining specimens. Project Genie is aimed at paid Google AI Ultra subscribers; Google has not announced a public API model ID, token pricing schedule, or third-party benchmark scores for Genie 4.