source: google ai: 11 demos of gemini omni and gemini 3.5 in action
level: business
gemini omni is a new model that can create video from mixed inputs like images, audio, and text. it lets users edit videos through natural language conversation, keeping characters and scenes consistent across multiple edits. demos show changing a sculpture to bubbles, dimming lights and adding recursive glass spheres, and transforming a violinist scene by moving the setting, hiding the violin, and shifting the camera angle.
gemini 3.5 flash is built for complex, long-horizon agentic workflows. it matches large model intelligence at flash speeds. one demo, powered by antigravity, shows it automatically renaming and categorizing unstructured files. another demo on ai studio generates different checkout flow ux designs in 60 seconds. it also powers new search features like information agents that run in the background and send updates on topics like sneaker releases.
the models are rolling out across google products. gemini omni flash is available to google ai subscribers and free on youtube shorts and create app. gemini 3.5 flash is the default for the gemini app and ai mode in search, and is available via antigravity, api, and enterprise platforms. upcoming features include custom search experiences and gemini spark, a personal agent that works across workspace tools.
why it matters: these models bring advanced video generation and agentic ai to everyday tools, enabling faster content creation and automated complex tasks for developers and users.
source: google ai: 11 demos of gemini omni and gemini 3.5 in action