Google's Gemini Omni Shows 2 Video Demos as I/O 2026 Nears

4 articles · Updated · 9to5Google · May 11

Google’s unannounced Gemini Omni surfaced for at least one user as a new video-generation tool that can remix clips, edit directly in chat and use templates.
Two demo prompts produced notably realistic results: a professor writing trigonometric proofs on a chalkboard and two men calmly eating spaghetti at a seaside restaurant.
Metadata suggests Omni may extend Google’s Veo model, though Google has not clarified how the new feature fits into Gemini’s broader video strategy.
An AI Pro account’s usage tab showed those two prompts consumed 86% of daily usage, hinting at meaningful limits or high compute costs for video generation.
Google has yet to announce Omni officially, but with I/O 2026 close and the company recently saying “video’s here to stay,” more details may arrive soon.

When AI can generate flawless video, what unique skills will future human creators need to succeed?

As AI perfects fake reality, what is Google's plan to stop a wave of undetectable deepfakes?

Is the high cost of AI video creating a new digital divide for aspiring filmmakers?