Google I/O 2026: Gemini 3.5 Flash & Omni Debut

Google used its I/O 2026 keynote to roll out more than 100 announcements, anchored by three headline launches: Gemini 3.5 Flash, a new Gemini Omni content-generation model, and what Google calls the biggest redesign of its Search box in over 25 years. The event's stated focus was "making AI helpful for everyone in new ways," and Google notably used Gemini and other AI tools to produce parts of the keynote itself, including a short film, a generative visual identity, and AI-generated title cards for speakers.

The headline launches

Gemini 3.5 Flash is now generally available today through Google Antigravity, the Gemini API, Google AI Studio and Android Studio, and is rolling out globally in the Gemini app. (Sources differ on the full platform list — one report adds Gemini Enterprise Agent Platform and Gemini Enterprise to this rollout.) Google describes it as the first in a new model series combining "frontier intelligence with action," built for agentic and coding tasks over complex, long-horizon workflows. On benchmarks, Gemini 3.5 Flash reportedly outperformed Gemini 3.1 Pro, scoring 76.2% on Terminal-Bench 2.1, 1656 Elo on GDPval-AA, and 83.6% on MCP Atlas. (Benchmark methodology for these tests was not detailed in available materials.)

Gemini 3.5 Pro, currently used internally at Google, is set to roll out next month.

Gemini Omni, a new model capable of generating content — starting with video — from any input type, is also making its debut. Its first release, Gemini Omni Flash, is rolling out today to all Google AI Plus, Pro and Ultra subscribers globally via the Gemini app and Google Flow, and to YouTube Shorts Remix and YouTube Create app users aged 18+ at no cost. Videos generated by Omni carry an imperceptible SynthID watermark — a detail that signals Google's continued attention to AI-content authenticity as generative video becomes more accessible.

Search gets an AI overhaul

Google is repositioning Search around AI with a new, seamless AI Overviews + AI Mode experience that went live today worldwide on desktop and mobile. AI Mode itself has now surpassed 1 billion monthly users, and its underlying default model is being upgraded globally to Gemini 3.5 Flash.

Several new capabilities are set to follow this summer:

Search agents, letting users create and manage multiple AI agents for tasks directly within Search.
Information agents, which operate 24/7 in the background to keep users updated on topics, tasks or projects — rolling out first to Google AI Pro and Ultra subscribers.
Generative UI, built with Google Antigravity and Gemini 3.5 Flash's agentic coding capabilities, enabling dynamic layouts and custom mini-app experiences in Search — free for everyone.
Custom Antigravity experiences (e.g., mini apps) in Search, rolling out to subscribers first — though sources differ on whether this means subscribers generally or specifically Google AI Pro and Ultra subscribers in the U.S.
Universal Cart, a shopping cart that works across merchants and services, launching in the U.S. across Search and the Gemini app, with YouTube and Gmail integration to follow. Details on payment handling, privacy, and merchant integration mechanics were not disclosed.

Personal Intelligence in AI Mode is also expanding to nearly 200 countries and territories across 98 languages, with no subscription required.

A redesigned Gemini app and new hardware

The Gemini app is getting a design overhaul called "Neural Expressive," featuring fluid animations, new typography and haptic feedback, rolling out across Android, iOS and web. It also introduces Gemini Spark, a 24/7 personal AI agent that can take action on users' behalf, coming to the Gemini desktop (macOS) app this summer, alongside new voice features later in the season. A related agent, Daily Brief, will deliver personalized morning summaries pulled from Gmail, Calendar and other sources, rolling out to Google AI subscribers 18+ starting in the U.S.

On the hardware side, Google announced intelligent eyewear as the next chapter for Android XR, split into two categories — audio glasses and display glasses. Only the audio glasses have a confirmed timeline, launching later this fall; no details on display glasses or pricing for any of the new hardware or models were provided.

Why founders should care

The scale and simultaneity of these launches likely signals that Google is positioning Gemini as a horizontal platform rather than a single point of integration — founders building AI-dependent products may want to evaluate multiple entry points (API, Antigravity, AI Studio, Search agents) rather than betting on just one.

The staggered rollout schedule — some features live today, others arriving "next month," "this summer," or "later this fall" — suggests founders relying on these capabilities should plan phased release timelines rather than assuming immediate global availability. This is particularly relevant given that several features (information agents, custom Antigravity experiences) are launching first to Pro/Ultra subscribers or U.S. users, which could mean delayed feature parity for teams targeting international or free-tier audiences.

The benchmark gains reported for Gemini 3.5 Flash on agentic and coding tasks may inform build-vs-buy decisions for startups developing coding assistants or agent-based tools, though the underlying methodology wasn't disclosed, so these figures should be weighed cautiously.

The emergence of persistent, always-on agents — information agents, Gemini Spark, Daily Brief — could indicate a shift in user expectations toward continuous AI assistance, a trend product teams may want to factor into their own roadmaps. At the same time, this rapid rollout of autonomous agents across Search and Gemini may raise data privacy and control questions that enterprise-focused founders will need to account for.

New surfaces like Search agents, generative UI, and Universal Cart could open integration opportunities for startups building mini-apps or e-commerce tools within Search, while the entry of Android XR eyewear signals an early-stage hardware/software opportunity for wearable AI developers — though as an unproven category at scale, timelines for third-party integration remain uncertain.

What's still unclear

Several important details were not disclosed in available materials: pricing for Gemini 3.5 Flash, Gemini 3.5 Pro, Gemini Omni, and the new eyewear; the scope of "trusted tester" or Beta access to Gemini Spark; and how these announcements compare competitively to offerings from other AI platforms. Founders evaluating these tools should watch for pricing and access details as they become available.