Google launches Gemini Omni capable of chat-based video generation, editing

Tech giant Google LLC on Wednesday launched Gemini Omni, a groundbreaking artificial intelligence (AI) model family capable of generating videos from most common inputs and allowing them to be edited through conversation.

Gemini Omni is among the key innovations in AI announced by the company at the Google I/O 2026 event, where it also launched the Gemini 3.5 Flash model—its most powerful agentic and coding model yet—and intelligent eyewear, among other innovations.

Google said that Gemini Omni follows the success of Nano Banana, its advanced AI image generation and editing model that can create images from various inputs, but applies that technology to video.

“With Omni, you can combine images, audio, video and text as input and generate high-quality videos grounded in Gemini’s real-world knowledge,” Google said in a press statement.

“You can also easily edit your videos through conversation,” it added.

The first model of the Omni family, Gemini Omni Flash, was the first to roll out and is available on the Gemini app, Google Flow, and YouTube Shorts.

It will also be rolled out to developers and enterprise customers via APIs (application programming interface) in the coming weeks.

Currently, it only creates video, but Google noted that in time, the model will also support output modalities like images and audio.

Google recommended using the model to create videos with the user’s voice by using avatars, which generate a digital version of a person that looks and sounds like them.

It also said that it is working on a capability where the audio and speech of a video can be changed using the model. However, it noted that this is still being tested and is in the process of identifying how to bring the capability to users responsibly.

The tech giant emphasized that videos generated with Omni contain SynthID, its proprietary digital watermark used to easily detect AI content made by Gemini Omni. — Edg Adrian A. Eva