Google Launches Gemini Omni, Its New AI Video Generator
Gemini Omni generates AI videos from text prompts, images, and user uploads.
Tech company Google launched Gemini Omni, an artificial intelligence (AI) multimodal model family designed to generate video content.
Announced during Google I/O 2026, an annual conference hosted by the technology giant, Omni improves on the company’s past offerings in terms of the physics displayed in videos.
The introduction of Gemini Omni expands the Gemini product lineup, which includes the AI image generator Nano Banana.
“It’s a step change in simulating things like kinetic energy and gravity. Previous systems would have found these concepts difficult,” said Demis Hassabis, Google DeepMind co-founder and chief executive officer, during his presentation of Omni at the event.
Google DeepMind, a subsidiary of Google’s parent company Alphabet Inc., is a research laboratory focused on AI.
According to Hassabis, Omni can translate “complex ideas into highly accurate videos.”
With it, users can generate videos by providing prompts to base the generation on. Videos can also be created by uploading images that Omni can use as references when generating.
Users can edit their own videos, upload them, add additional elements, and/or change the art style through conversations with Omni.
The new video generation offering marks the first step into Google’s main goal for the multimodal model family, to create “any output from any input.”
“This was always our goal with Gemini and why we built it to be multimodal from the very start,” Hassabis said.
Google Doubles Down on AI
Google joins tech companies like Apple, Microsoft, Amazon, and Meta in investing in artificial intelligence research, as well as adapting AI into their offerings.
In April 2026, Google announced its investment in Anthropic, the makers behind the Claude chatbot, valued at $40 billion. Google committed $10 billion in initial cash and said it will provide an additional $30 billion should Anthropic meet performance targets.
Google also credits the AI plans for several of its products and services as a boost in its recent quarterly earnings.
Alphabet Inc. posted its consolidated revenue earnings at $109.9 billion, a 22% increase from the previous quarter. The recent number marks the company’s 11th consecutive quarter of double-digit growth.
Alphabet and Google CEO Sundar Pichai said in an April 29, 2026, call with investors that the positive response to AI’s inclusion in its offerings drove overall growth.
“It’s clear that our AI investments and full-stack approach are driving performance across our business,” said Pichai.
Frequently Asked Questions
Unveiled at Google I/O 2026, Gemini Omni is a next-generation multimodal model family designed to generate and edit high-quality video content. What sets it apart from previous text-to-video tools is its grounding as a world reasoning model that inherently understands complex physical properties. It simulates forces like gravity, kinetic energy, and fluid dynamics, allowing it to translate descriptions into lifelike videos that bridge the gap between photorealism and accurate storytelling.
Gemini Omni treats an uploaded or generated video as a fluid starting point rather than a finalized asset. Through conversational multi-turn editing, users can continuously alter specific details of a clip over the course of a single chat. Users can prompt the AI to change the environment, shift camera angles, modify the overall art style, swap out background scenery, or seamlessly add and remove characters and objects while preserving character consistency across frames.
Reflecting Google’s foundational goal of generating “any output from any input,” the Omni architecture natively accepts a diverse range of overlapping inputs. Users can provide standard natural language text prompts, upload reference images to animate or establish a specific style, or layer in audio assets like voiceovers and soundtracks. It can even take pre-existing video clips to handle complex style transfers, continuations, or detailed scene modifications.
Google launched the first model of this family, Gemini Omni Flash, making it immediately available to paid Google AI Plus, Pro, and Ultra subscribers within the Gemini app and Google Flow. To ensure the tools are widely accessible, Google is also rolling out Gemini Omni Flash globally within YouTube Shorts and the YouTube Create app, allowing creators to utilize advanced video generation on their mobile devices at no extra cost.
The most frequent mistake creators make is viewing Gemini Omni as merely a flashy text-to-video prompt tool like Google’s older Veo model, rather than an integrated workflow engine. Because Omni shares the same reasoning infrastructure as Google’s flagship language models, it can be plugged into broader automation systems—such as automatically generating marketing videos and digital avatars from text-based product descriptions in a CMS.
