Key Takeaways
- Gemini 3.5 Flash is now the default model across Google's apps and Search
- Gemini Spark is Google's answer to OpenClaw, an always-on background agent
- Omni Flash can generate video from text, photos, video, and audio inputs
Google's I/O 2026 keynote on May 19th was packed with AI announcements. CEO Sundar Pichai walked through updates to the Gemini model family, new agentic capabilities, smart glasses progress, and tools that let developers build Android apps through prompts alone.
If you missed the two-hour livestream, here's a breakdown of the 13 biggest announcements from the event.
Gemini 3.5 Flash and Pro
Google launched Gemini 3.5 Flash as its new default model for the Gemini app and AI Mode in Search. The company says it's significantly faster than its predecessor, handles agentic tasks better, and generates richer web UIs and graphics. Improved guardrails should reduce harmful content generation while cutting down on false positives that flagged safe queries as unsafe.
Gemini 3.5 Pro arrives next month with additional capabilities for more complex reasoning tasks.
The Gemini App Gets a Visual Overhaul
Google introduced a redesign for the Gemini app called "neural expressive." The new look includes updated animations, color accents, a new font, and haptic feedback. It started rolling out May 19th on web, Android, and iOS.
Gemini Omni: Multimodal Generation
Beyond the 3.5 updates, Google announced Gemini Omni, a new family of models built for multimodal generation. The first model, Omni Flash, is available now in the Gemini app, Google Flow, and YouTube Shorts.
Unlike Google's Veo model, which only accepts text prompts, Omni Flash generates video clips from a mix of inputs: text, photos, video, and audio. Google says future versions will be able to "create anything from any input."
Gemini Spark: Google's Always-On Agent
Gemini Spark is Google's response to OpenClaw. It's an always-on AI agent that runs in the background on Google Cloud virtual machines around the clock. Spark can write emails, create study guides, and monitor for hidden credit card fees without manual prompting.
Powered by Gemini 3.5 Flash, Spark connects to Google Workspace apps including Docs, Gmail, Sheets, and Slides. Third-party integrations with Canva and Instacart are also available at launch. Google plans to add local file access through the Gemini app on macOS.
Step-by-step guide for setting up and using Google's new agent capabilities
Vibe-Code Android Apps in AI Studio
Google AI Studio now lets users build full native Android apps through prompts. Instead of writing code line by line, developers describe what they want and the system generates a working app. This marks Google's biggest move yet into prompt-driven development.
Project Aura Smart Glasses
Google provided updates on Project Aura, its smart glasses initiative. The company is partnering with Warby Parker on hardware, bringing AI-powered features to stylish frames designed for everyday wear.
AI Shopping Features and Universal Cart
Google showed off AI-powered shopping features, including a Universal Cart that works across retailers. The goal is to let users add items from different stores to a single checkout experience, with AI handling product matching and price comparisons.
Google Beam: Video Agents
HP's Google Beam device appeared during the keynote, showing an AI agent displayed on screen for video interactions. This points to a future where AI assistants have a visual presence in home and office settings.
Search and Gmail Updates
Search gets deeper AI integration with Gemini 3.5 Flash powering the AI Mode. Gmail receives new features as well, though Google didn't detail all the specifics during the keynote.
Our full technical breakdown of the Gemini updates
What This Means for Developers
The announcements signal Google's push toward agentic AI that works autonomously. Spark running 24/7 on cloud VMs is a clear challenge to OpenClaw's model. The prompt-based Android development in AI Studio could reshape how smaller teams build mobile apps.
For enterprises, the Workspace integrations with Spark matter most. An agent that monitors emails for hidden fees or drafts responses could save hours per week for knowledge workers.
Logicity's Take
Frequently Asked Questions
When is Gemini 3.5 Pro releasing?
Google said Gemini 3.5 Pro will launch next month, following the May 19th release of Gemini 3.5 Flash.
What can Gemini Spark do?
Spark is an always-on AI agent that writes emails, creates study guides, monitors for hidden fees, and connects to Google Workspace and third-party apps like Canva and Instacart.
How is Gemini Omni different from Veo?
Veo only generates video from text prompts. Omni Flash can generate video from text, photos, video, and audio inputs combined.
What is Google's neural expressive design?
It's the new visual design language for the Gemini app, featuring updated animations, colors, fonts, and haptic feedback.
Can I build Android apps with AI prompts now?
Yes, Google AI Studio now supports prompt-based development of native Android applications.
Need Help Implementing This?
Gemini Spark: Rollout Timeline and Third-Party Integrations
The new article provides specific rollout details, including testing this week and a beta launch for US AI Ultra users next week. It also lists third-party integrations with Canva, OpenTable, and Instacart, and highlights upcoming features like browser operation and a desktop app for local file tasks.
Advanced Creative Control with Gemini Omni
The new article provides specific details on Gemini Omni's distribution through Google Flow and YouTube Shorts, as well as its ability to perform conversational video editing with character consistency. It also highlights the model's advanced understanding of physical forces like gravity and its initial limitation of supporting only voice references for audio output.
Huma Shazia
Senior AI & Tech Writer
Produced with AI assistance and reviewed by the Logicity editorial team. Learn more in our Editorial Policy.
