Google I/O 2026: Gemini Spark, Omni, and Every Major Announcement

Google I/O 2026 was the company's biggest AI showcase yet — and for once, the hype matched the reality. In a single keynote, Google announced three new Gemini model families, a personal AI agent that can act on your behalf, a multimodal video generation model, enterprise-grade managed agents, and sweeping upgrades across every major product it ships.

Here is a breakdown of everything that actually matters.

The New Gemini Models

Gemini 3.5 Flash

The headline model launch is Gemini 3.5 Flash — the first in what Google is calling its "frontier intelligence meets action" series. The key claim: it delivers intelligence that rivals large flagship models while maintaining the speed users expect from the Flash series. In benchmark testing, it outperforms Gemini 3.1 Pro on challenging coding and agentic tasks.

Gemini 3.5 Pro, the full-power version, is currently in testing and set to roll out in June 2026.

Gemini Omni

Gemini Omni is built for creation, not just conversation. It accepts image, audio, video, and text as input, then generates video output grounded in real-world knowledge. Think of it as Google's answer to Sora and Runway — but woven directly into the Google ecosystem, with easy editing built in from the start.

Gemini Spark

This is the one most users will care about day-to-day. Gemini Spark is Google's "personal agent" — it doesn't just answer questions, it takes actions on your behalf across your digital life. Booking appointments, managing emails, making purchases, navigating apps. Spark is rolling out the week of June 1 to Google AI Ultra subscribers in the US first.

Managed Agents in the Gemini API

Developers got a significant new tool: Managed Agents in the Gemini API. A single API call now provisions a remote Linux environment where an agent can reason, plan, call tools, execute code, manage files in an isolated sandbox, and browse the web — all without you needing to spin up your own infrastructure. It is a direct shot at OpenAI's Codex and Anthropic's Claude Code for AI-assisted development workflows.

Android XR and Smart Glasses

Google announced new Android XR glasses built in partnership with Qualcomm. The glasses run a lightweight version of Gemini on-device, overlay contextual information into your field of view, and are designed to integrate with your phone rather than replace it. No release date or pricing was given at the keynote, but prototype footage was shown and the reception from the crowd was notably enthusiastic.

Google Search Gets Smarter

Search is getting deeper Gemini integration throughout. AI Overviews are expanding, and Google is testing AI-generated interfaces that dynamically restructure the search results page based on what it thinks you are trying to accomplish — not just showing links but composing answers with sources.

SynthID, Google's AI content watermarking system, has now been used over 50 million times globally across images, video, and audio. It is expanding into Search and Chrome, making it easier to identify AI-generated content while browsing.

YouTube and Workspace

YouTube is getting Gemini-powered video summaries, auto-generated chapters, and smarter search. On the Workspace side, Gmail, Docs, and Sheets are all receiving deeper agent-style features — including AI that can draft full documents, respond to email threads, and analyze spreadsheet data on command.

Google Shopping and Privacy

Google announced significant improvements to its shopping experience powered by AI. The system can compare products across retailers, surface price-drop alerts, and even complete purchases through Gemini Spark. If you plan to explore any of these shopping tools or try out new Google services, consider using a temporary email at app.fasttempmail.com before connecting your real inbox — retail platforms are notorious for flooding new users with marketing as soon as they have an address on file.

What Is Still Missing

For all the ambition on display, a few things were notably absent. Pricing was not announced for Gemini Spark beyond the requirement of an existing Ultra subscription. Gemini 3.5 Pro remains in closed testing. And the Android XR glasses are still prototype territory — there is no indication of when consumers will actually be able to buy them.

Google also did not address ongoing concerns about privacy in agentic AI. When an AI agent acts on your behalf across the web, it accesses a significant amount of personal data. That conversation is still waiting to happen in a meaningful way.

FAQ

What is Gemini Spark and when is it available? Gemini Spark is Google's personal AI agent that can take actions on your behalf — booking appointments, managing emails, and navigating apps. It launches the week of June 1, 2026, initially for Google AI Ultra subscribers in the US.

What is the difference between Gemini 3.5 Flash and Gemini Omni? Gemini 3.5 Flash is Google's fastest, most capable reasoning model for coding and agentic tasks. Gemini Omni is a separate creation-focused model that generates video from multimodal inputs including text, image, and audio. They serve different use cases.

What are Android XR glasses? Android XR glasses are a new hardware category Google is developing with Qualcomm. They are lightweight smart glasses that run on-device Gemini AI and overlay contextual information into your field of vision. They are still in prototype stage as of I/O 2026.

What is SynthID? SynthID is Google's AI content watermarking system that embeds imperceptible signals into AI-generated images, video, and audio. It has been used over 50 million times and is expanding into Chrome and Google Search in 2026.