Google: Nine Demo Videos Showcase Gemini Omni and 3.5 Capabilities

Google released nine demo videos showcasing the capabilities of the Gemini Omni and Gemini 3.5 models, unveiled at the I/O 2026 conference. Three videos are dedicated to Gemini Omni, and six to the Gemini 3.5 Flash model.

What does Gemini Omni show?

The Gemini Omni demos focus on editing video through conversation. In one example, a user changes content using natural language, with “characters staying consistent, physics convincing, and the scene remembering what came before” — for instance, a statue turning into bubbles. Another demo shows “reimagining the action,” where characters or objects are added, and a third shows iterative editing through multiple instructions without losing the context of the original scene.

What does Gemini 3.5 Flash bring?

The six Gemini 3.5 Flash demos emphasize agentic and generative abilities. One shows agentic task execution in which the model “performs multi-step workflows to automatically rename and categorize unstructured materials” using the Antigravity framework. Another generates different UX approaches for checkout inside AI Studio, and a third shows information agents in search that “intelligently reason through information.”

What is a generative user interface?

A generative interface (generative UI) means that search builds a custom visual interface, including visual tools and simulations, tailored to the user’s needs. In addition, the demos also show custom dashboards — personal companions and mini-applications for tasks such as wedding planning or fitness tracking.

What is Gemini Spark?

The final demo introduces Gemini Spark, a personal AI agent that runs continuously (24/7) and is “deeply integrated with Workspace tools” such as Gmail and Docs, performing tasks on the user’s behalf. Together, the set of demos outlines Google’s vision in which models move from answering queries to proactively executing workflows.