Google appears set to boost Gemini's video generation features via a model named Gemini Omni. Recent leaks indicate users could soon generate and edit AI videos right inside Gemini. Preliminary demonstrations reveal superior motion realism, sharper text display, and enhanced scene arrangement. While not yet official, the timing before Google I/O 2026 hints at upcoming announcements for AI-driven video tools.

Google Gemini Omni Set to Emerge as Cutting-Edge AI Video Generator

A 9to5Google report notes that one Gemini user encountered a prompt labeled "Create with Gemini Omni." Google describes it as a fresh video generation system capable of remixing clips, editing in chats, using templates, and handling various creative functions.

Details on Omni's place in Google's overall AI video roadmap remain vague. Feature metadata points to it being developed atop the company's Veo platform, likely extending rather than supplanting it.

A highlighted leak featured a prompt for a professor to write and detail trigonometric proofs on a chalkboard. The output video managed equations with notable precision, alongside lifelike hand gestures and facial reactions - outperforming many current AI video generators.

Another test recreated the classic "spaghetti test," a standard for assessing video AI prowess. It depicted two men at a coastal eatery, twirling and eating spaghetti amid conversation. Reports claim Gemini Omni delivered far more believable results, with refined hand-eye coordination and fluid object handling.

The tester also spotted a new usage tracker in Gemini. Generating just two videos reportedly drained 86 percent of the daily limit on the Google AI Pro subscription, signaling tighter caps for video tasks compared to text or images.

Google has emphasized video generation in its AI roadmap. As Google I/O 2026 approaches, it may unveil Gemini Omni there, clarifying its integration with the wider Gemini lineup. No confirmation exists yet, so view these leaks skeptically pending official word.