Google I/O 2026: Gemini 3.5 and Gemini Omni promise a leap in AI capabilities

Google presented at its annual Google I/O 2026 conference a series of significant AI announcements, led by Gemini 3.5, the new version of its AI assistant with improved reasoning and action capabilities, and Gemini Omni, a completely new family of multimodal models.

Gemini 3.5

The new version of Google's main model significantly improves logical reasoning and the ability to complete complex actions. The company claims Gemini 3.5 can handle multi-step tasks with greater accuracy, integrating with external tools more fluidly than previous versions.

Gemini Omni

Gemini Omni represents an architecture shift: a family of models designed from the ground up to process text, audio, video, and images simultaneously and coherently. Unlike models that were adapted for multimodality, Omni was born with that capability as a foundation.

Combating deepfakes

Among the announcements are improved tools to identify AI-generated images and videos, a direct response to the growing deepfake problem. Google said these tools will be available to developers through its API.

Chrome integration

Google also announced that Gemini will integrate more deeply with Chrome, allowing the AI assistant to help with navigation tasks, form completion, and information retrieval autonomously within the browser.

Design changes

Beyond technical capabilities, Google redesigned the visual interface of several products to integrate Gemini more prominently, consolidating its AI assistant as the center of the Google experience.