Google I/O 2026: Everything Google Announced — AI Models, Smart Glasses, Gemini Agents & More
Google I/O 2026 was one of the most packed developer and product conferences in the company’s history. Held on May 19, 2026, at the Shoreline Amphitheatre in Mountain View, California, Google CEO Sundar Pichai took the stage to announce over 100 new products, features, and tools — all united under a single bold theme: the agentic Gemini era.
From next-generation AI models and a completely redesigned Google Search to AI-native laptops, smart glasses, and autonomous background agents, Google I/O 2026 signaled a fundamental shift in how Google intends to make AI useful in everyday life — not just as a chatbot you talk to, but as a system of agents that act on your behalf.
1. New Gemini AI Models
Gemini 3.5 Flash — The New Default
The headline model launch was Gemini 3.5 Flash — the first in the new Gemini 3.5 series. It outperforms Gemini 3.1 Pro on nearly all benchmarks while running four times faster than rival frontier models. Scores: 76.2% on Terminal-Bench 2.1, 1,656 Elo on GDPval-AA, and 83.6% on MCP Atlas. Available now via the Gemini app, Google AI Studio, and the Gemini API.
A more powerful Gemini 3.5 Pro is in the pipeline, scheduled for June 2026.
Gemini Omni — The Multimodal World Model
Gemini Omni processes and generates text, images, video, and audio within a single unified architecture — no modality-switching. Its standout ability is generating high-quality video grounded in Gemini’s knowledge of physics, science, history, and culture.
Gemini Omni Flash launched at I/O and is available free inside YouTube Shorts Remix and the YouTube Create app (18+). All generated videos carry an imperceptible SynthID watermark.
Gemini Spark — The On-Device Small Model
Gemini Spark runs entirely on-device on Android and ChromeOS — no internet required. Its “Adaptive Context” feature adjusts processing based on available memory. It becomes the default inference layer for Pixel devices and also powers the new Gemini Spark personal agent.
2. Google Search — Biggest Redesign in 25 Years
Google called this the biggest upgrade to its Search box in over 25 years. The redesigned Search now accepts multimodal inputs: text, images, files, videos, and open Chrome tabs. AI Overviews and AI Mode have merged into one seamless experience — already used by 1 billion+ monthly users.
- Agentic Search — Background agents monitor blogs, news, and real-time data 24/7 on your behalf (summer 2026, Pro/Ultra)
- Generative UI — Search assembles custom interactive layouts (graphs, tables, simulations) for complex queries (free, summer 2026)
- Personal Intelligence — Connect Gmail, Google Photos, and Calendar for personalized results, expanding to 200 countries in 98 languages at no cost
3. The Gemini App: Agents & New Identity
Neural Expressive Redesign
Complete visual overhaul with fluid animations, vibrant colors, new typography, and haptic feedback. Gemini Live is now integrated inline with a faster model. Regional dialect accents coming soon.
Gemini Spark Agent — 24/7 AI Assistant
A 24/7 personal AI agent that works in the background even when your device is off. Running on Google Cloud via Antigravity, it can:
- Scrape and organize inboxes
- Synthesize notes and draft documents
- Execute multi-step workflows
- Integrate with 30+ third-party apps (Uber, OpenTable, Lyft, Zillow)
Exclusive to AI Ultra subscribers in the U.S. at launch, in beta the week after I/O.
Daily Brief
A pre-configured morning digest agent. Works overnight analyzing your inbox, calendar, and tasks — surfaces the most important items and suggests next steps. Live for all AI subscribers (18+) in the U.S. from May 19, 2026.
4. Google Workspace Updates
Gmail Live
Talk to your inbox out loud — ask about flight details, school events, or deadlines without digging through threads. Rolling out to AI Pro and Ultra subscribers, summer 2026.
AI Inbox (personalized draft replies) is also expanding from Ultra-only to AI Plus and Pro subscribers in the U.S.
Docs Live
Fully voice-driven document creation — dictate thoughts and Gemini organizes them into a formatted document, with context pulled from Gmail, Drive, Chat, and the web. Rolling out summer 2026, Pro and Ultra.
Talk to Keep
Brain-dump verbally while Keep auto-organizes your stream of consciousness into structured notes and lists in the background. Summer 2026, Pro and Ultra.
Google Pics
A Canva-like AI image editor built on Google’s Nano Banana image model. Supports object segmentation, text editing, text translation, and full Docs/Slides integration. Limited Trusted Tester launch May 19; global rollout summer 2026.
5. Universal Cart — AI Shopping Hub
Universal Cart lets users add items from Search, the Gemini app, YouTube, and Gmail into one persistent AI-powered cart. Background agents then:
- Track price drops
- Monitor stock availability
- Check product compatibility
- Surface relevant deals and loyalty perks
Built on Google Wallet and the Agent Payments Protocol (AP2) — which will eventually let AI agents complete purchases autonomously. Rolling out across Search and Gemini in summer 2026.
6. Google Antigravity 2.0 — Developer Platform
A major evolution of Google’s agent-first developer platform:
| Tool | What It Does |
|---|---|
| Antigravity Desktop App | Orchestrate multiple parallel agents from one hub |
| Antigravity CLI | Build agents from the terminal, no GUI needed |
| Antigravity SDK | Access the same agent harness powering Google’s products |
| Managed Agents API | One API call spins up a cloud Linux sandbox for agent execution |
| Subagent Teamwork | Distribute multi-day tasks across parallel subagents (research preview) |
Also announced: Build with Gemini XPRIZE Hackathon — a global competition with a $2 million prize pool, the largest in hackathon history.
7. Google Flow — Creative Studio Upgrades
- Gemini Omni Flash in Flow — Available to all AI subscribers globally for video projects
- Flow Agent — Multi-step creative tasks, batch editing, and asset organization
- Flow Tools — Build custom video effects and animations via natural language (no coding required), shareable with the community
- Flow Music — Conversational music video direction with control over lyrics, genre, and instruments
- Pomelli — Brand content and website design
- Stitch — Real-time collaborative design with voice/text steering and codebase import
8. Gemini for Science
A suite of AI research tools on Google Labs:
- Hypothesis Generation — Multi-agent “idea tournament” using Co-Scientist to simulate the scientific method
- Computational Discovery — Uses AlphaEvolve and ERA to generate and score thousands of code variants in parallel
- Literature Insights — NotebookLM-powered scientific literature search with side-by-side comparison tables
- Science Skills — Integrates 30+ life science databases including UniProt, AlphaFold Database, and InterPro
9. Android XR Smart Glasses — Fall 2026
Developed with Samsung, Android XR intelligent eyewear comes in two forms:
| Type | Features |
|---|---|
| Audio Glasses | Gemini voice assistance, directions, texts, photo capture |
| Display Glasses | Information overlaid in the field of view |
First audio glasses arrive fall 2026 in frame collections from Gentle Monster and Warby Parker.
10. Googlebook — AI-Native Laptop
Googlebook is the official Chromebook successor — AI-native laptops running a merged Android + ChromeOS OS built around Gemini. First models from Acer, ASUS, Dell, HP, and Lenovo in fall 2026.
Key features:
- Magic Pointer — AI cursor that surfaces Gemini suggestions on hover
- Full Android app compatibility
- Custom AI widgets via natural language
- Seamless phone-to-laptop mirroring
11. Android 17 & Gemini Intelligence
Android 17 (announced May 12 at the Android Show: I/O Edition) introduces the Gemini Intelligence brand for all AI-powered Android features on flagship devices.
| Feature | Details |
|---|---|
| AI-Generated Widgets | Gemini creates personalized home screen widgets |
| Chrome Auto Browse | Gemini executes web tasks autonomously (June 2026) |
| AI-Powered Autofill | Fills passports, license plates, complex forms across apps |
| Pause Point | Digital Wellbeing labels for distracting apps |
| 3D Emoji Refresh | 4,000+ emojis redesigned in 3D |
| Android Auto + Gemini | Gemini in cars for texts, navigation, and news |
12. SynthID Watermarking Expansion
Google’s imperceptible AI watermarking technology made a major leap at I/O 2026:
- Now being integrated into Google Search, Chrome, and Circle to Search
- New SynthID Detector portal scans uploaded content (images, audio, video, text) for AI generation
- C2PA Content Credentials added to verify image authenticity
- OpenAI, Kakao, and ElevenLabs signed on to bring SynthID to their platforms
- Over 50 million uses in the Gemini app; over 100 billion images and videos watermarked globally
13. Ask YouTube
A new conversational search mode for YouTube. Ask complex natural language questions and Ask YouTube compiles relevant videos and Shorts, jumping to the exact moments that answer your question. Rolling out May 2026 for English desktop users in the U.S.
14. New Google AI Subscription Pricing
Google restructured its AI tiers at I/O 2026:
| Plan | Monthly Price | Key Inclusions |
|---|---|---|
| AI Plus | $7.99 | Gemini 3.5 Flash, YouTube Premium Lite |
| AI Pro | $19.99 | YouTube Premium Lite, expanded features |
| AI Ultra | $99.99 (new entry) | 5x usage vs. Pro, 20TB storage, Gemini Spark beta |
| AI Ultra (legacy) | $200 (down from $250) | Full existing Ultra features |
Usage now measured in compute consumed (not prompts) and refreshes every 5 hours instead of daily.
Key Takeaways
The agentic shift is real. Every major product — Gemini Spark Agent, Daily Brief, Agentic Search, Universal Cart, Gmail Live — shares the same philosophy: AI should act for you, not just respond to you.
Google built a unified stack. Gemini 3.5 Flash (reasoning) + Antigravity (orchestration) + Managed Agents (execution) + Gemini Spark (on-device) + WebMCP (open web standard) form a single interconnected infrastructure across all Google products.
Hardware is back. Googlebook replaces Chromebook and Android XR glasses arrive in fall 2026 — both with Gemini as the primary interface, not the OS.
Safety at scale. SynthID’s expansion into Search and Chrome, with OpenAI signing on, marks AI watermarking becoming an industry standard.
Google I/O 2026 took place May 19, 2026, Shoreline Amphitheatre, Mountain View, California. Availability and pricing are subject to change.