Google Gemini Omni Revolutionizes AI Video Creation

Creating images with artificial intelligence is no longer groundbreaking, but the real frontier lies in the ability to modify, provide continuity, and refine an initial idea into something more elaborate. In the realm of video, this challenge is considerably greater, involving movement, time, physics, and the crucial coherence of characters and settings. Google DeepMind has introduced Gemini Omni, an AI model designed to tackle these very issues and drastically simplify video editing.

The comparison made by Google DeepMind is telling: think of Gemini Omni like Nano Banana, but for video. Nano Banana, launched in August 2025, was an AI image generator that achieved massive popularity, attracting 13 million users in just four days and generating over 5 billion images. Now, Gemini Omni Flash, the first iteration of the Gemini Omni family, arrives with the ambition to bring that same versatility and scale to audiovisual content creation.

According to the company, Gemini Omni Flash is designed to create content from any type of input. This means users can combine images, audio, video, and text as a starting point to generate high-quality videos. The key lies in its ability to integrate Gemini's vast real-world knowledge, ensuring greater coherence and realism in its creations.

Video editing with Gemini Omni is envisioned not just as a tool for generating clips from scratch, but as an interactive system capable of working on an existing scene through detailed instructions. Google highlights the possibility of modifying specific elements or completely transforming a source video, adjusting aspects such as aesthetics, action, environment, camera angle, style, or specific details. A strong point is its promise to maintain character consistency and scene continuity, while also offering more believable physics.

The shared prompt examples illustrate Gemini Omni's potential. Instructions like “Make the sculpture out of bubbles” or “When the person touches the mirror, make the mirror ripple beautifully like liquid, and the person’s arm turns into reflective mirror material” demonstrate the granularity and control offered by the model. Even for complex explanations, such as a “Claymation explainer of protein folding, everything is made out of clay, no hands, stop motion, accurate,” Gemini Omni aims to deliver detailed and stylized results.

Regarding its availability, Google has announced that Gemini Omni Flash is now rolling out to subscribers of Google AI Plus, Pro, and Ultra via Gemini and Google Flow. Additionally, its free deployment on platforms like YouTube Shorts and the YouTube Create App will begin this week. However, in corporate tests, a limit of three videos per day was observed until a specific date, suggesting that Google is rationing access, given the high computational resource demand of AI video generation.

The arrival of Gemini Omni comes at a crucial time for AI video generation, where competitors like OpenAI with Sora have generated significant buzz. While Sora promised much, its availability and final results have been debated, with its website and app ceasing to be available in late April 2026, though its API will continue to operate until September. Gemini Omni aims to establish itself as a robust and accessible alternative, integrating into Google's ecosystem to democratize advanced audiovisual content creation.