At Google I/O 2026, Google introduced two new models: Gemini Omni and the Gemini 3.5 family. Gemini Omni is designed to create content from any input, starting with video. It allows users to combine images, audio, video, and text to generate high-quality videos grounded in real-world knowledge. Users can also edit videos through natural language, with edits building on previous instructions and maintaining scene consistency. The model can transform videos into entirely new scenarios, such as changing the environment or adding new characters. Gemini 3.5 Flash, part of the Gemini 3.5 family, is built for handling complex, agentic workflows. It delivers performance comparable to large flagship models while maintaining the speed of the Flash series. Powered by Antigravity, it can automate multi-step tasks like renaming and categorizing unstructured assets. 3.5 Flash is now the default model for the Gemini app and AI Mode in Search globally, enabling new features like information agents that provide real-time updates. These agents operate 24/7, offering insights on topics like athlete sneaker collabs. Search will also generate custom experiences, such as fitness trackers and dashboards, using 3.5 Flash. Gemini Spark, a personal AI agent running on Gemini 3.5, is available to Google AI Ultra rs in the U.S. and helps manage digital tasks like creating nut-free snack lists. Gemini Omni Flash is rolling out to Google AI Plus, Pro, and Ultra rs globally through the Gemini app and Google Flow. *Source: [googleai](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-omni-3-5-videos/)*