Alongside its new flagship language model, Google used Google I/O 2026 to introduce Gemini Omni, a model designed to generate content across modalities and grounded in the broader world knowledge of its Gemini system. The first release in the family, Gemini Omni Flash, starts with video generation and editing and is rolling out across consumer products.
A model that creates from any input
Google describes Omni as the point where Gemini's ability to reason meets its ability to create. The model can take images, audio, video and text as inputs and produce high-quality video output, with image and audio generation planned over time. Built on the same natively multimodal foundation as earlier Gemini models, Omni extends the company's generative work that began with image tools and now moves into moving pictures.
World knowledge and physics
What sets Omni apart, according to Google, is that it does not just assemble scenes that look real but reasons about what should plausibly happen next. The company says the model carries an improved intuitive understanding of forces such as gravity, kinetic energy and fluid dynamics, and draws on Gemini's knowledge of science, history and culture to connect language, imagery and meaning. That emphasis on modeling how environments change over time reflects a broader research direction in deep learning and embodied AI systems like Google DeepMind's work with Boston Dynamics' Spot robot.
Conversational editing and avatars
Omni lets users edit video through natural language, with each instruction building on the last while keeping characters consistent and scenes coherent across multiple turns. Google also introduced an Avatars feature, which creates a digital version of a user so they can generate clips that look and sound like them. The company said every video produced with Omni carries its imperceptible SynthID watermark, and that creations can be verified through the Gemini app, Gemini in Chrome and Google Search. The launch arrives the same week as Google's Gemini 3.5 Flash agentic and coding model.
Where to try it
Gemini Omni Flash is rolling out to Google AI Plus, Pro and Ultra subscribers globally through the Gemini app and Google Flow, and is also coming to YouTube Shorts and the YouTube Create app at no cost. Google said developer and enterprise access through APIs will follow in the coming weeks. The release underscores how quickly foundation-model competition is intensifying, echoed by hardware moves such as Alibaba's purpose-built AI chip for its Qwen models.
Reporting based on coverage from Google (The Keyword blog).