One of the leaked demonstrations used a prompt asking for a professor to write and explain trigonometric proofs on a chalkboard.
Gemini Omni may be build on Google's Veo model
Google may be preparing to expand Gemini's video generation capabilities with a new model called Gemini Omni. A fresh leak suggests the tool will let users create and edit AI-generated videos directly within Gemini. Early demos show the model producing more realistic motion, cleaner text rendering, and improved scene composition. The feature has not been announced officially, but its appearance ahead of Google I/O 2026 suggests Google could soon reveal new plans for AI-powered video creation.
According to a report by 9to5Google, at least one Gemini user recently saw a prompt titled "Create with Gemini Omni." Google reportedly described the feature as a new video generation model that can remix videos, edit them in chat, apply templates, and perform other creative tasks.
The publication states that it's currently unclear how Omni fits into Google's broader AI video strategy. Metadata associated with the feature suggests Gemini Omni may be built on top of Google's existing Veo technology rather than replacing it entirely.
One of the leaked demonstrations used a prompt asking for a professor to write and explain trigonometric proofs on a chalkboard. The resulting clip reportedly handled written equations more accurately than many existing AI video tools, while maintaining realistic hand movements and facial expressions.
A second demo recreated the well-known "spaghetti test," a benchmark often used to evaluate AI video quality. The prompt showed two men dining at a seaside restaurant and eating spaghetti while conversing. According to the report, Gemini Omni generated a much more convincing result than earlier AI models, with improved hand coordination and more natural object interactions.
The user who accessed Gemini Omni also noticed a new usage tab inside Gemini. Two video generations reportedly consumed 86 percent of the account's daily quota on the Google AI Pro plan, indicating that video generation may carry stricter usage limits than standard text or image features.
Google has already made it clear that video generation remains a key part of its AI plans. With Google I/O 2026 just around the corner, the company may use the event to officially introduce Gemini Omni and explain how it fits into the broader Gemini ecosystem. Google has not confirmed the new model yet, so the leaked details should be treated with caution until the company makes an official announcement.
Get your daily dose of tech news, reviews, and insights, in under 80 characters on Gadgets 360 Turbo. Connect with fellow tech lovers on our Forum. Follow us on X, Facebook, WhatsApp, Threads and Google News for instant updates. Catch all the action on our YouTube channel.