Gemini Omni Leak — Google Nearing Reveal of Integrated Video Model at May I/O

Gemini's video tab leaked a 'Powered by Omni' UI string on 2026-05-02. First demo dropped 2026-05-11. Expected unveil at Google I/O 2026 (May 19-20). Native video generation + chat-based editing in on

Developer Google DeepMind

Leak Date 2026-05-02 (X user @Thomas16937378)

First Demo 2026-05-11

Expected Official Announcement 2026-05-19 ~ 05-20 (Google I/O 2026 keynote)

Positioning Integrated Multimodal — Generate and edit text/image/audio/video from a single model.

How the Leak Unfolded

X user @Thomas16937378 discovered the UI string "Start with an idea or try a template. Powered by Omni" in Gemini's video generation tab, leading to a report by TestingCatalog. Subsequently, Google released its first demo video on May 11. A key differentiator compared to Sora, Veo, and Seedance is the ability to directly edit videos within a chat interface.

Why It Matters

- First top-tier foundation model with native video generation — generating naturally from text chat without a separate dedicated video model.

- Potential for integration with Google services like Workspace, YouTube, and Photos.

- Direct competition against Sora 2, Seedance 2.0, Kling, and others.

What to Watch For

Official launch at the May 19-20 keynote, and whether Gemini Advanced subscribers will receive immediate access.

Synopsis

The aforementioned content covers IT industry news and predictions related to the leak of an AI video generation model called 'Gemini Omni,' which Google is expected to unveil at May I/O. Therefore, it is not a storytelling work like a feature film or animation.

Maru