source: google ai: 100 things we announced at i/o 2026
level: technical
google launched gemini 3.5 flash, a fast frontier model for agentic tasks, available on google antigravity and ai studio. it outperforms gemini 3.1 pro on coding benchmarks and handles long-horizon tasks at lower cost. gemini 3.5 pro is coming next month. gemini omni generates video from any input, with physics understanding and synthid watermarking. omni flash is in the gemini app, google flow, and youtube shorts remix.
ai search now uses gemini 3.5 flash, with a redesigned search box accepting text, images, and files. ai mode and ai overviews merge into one experience. information agents will monitor topics and send updates, launching this summer. search will offer generative ui and custom mini apps via antigravity. personal intelligence expands to more countries without a subscription. universal cart tracks deals and suggests payment methods across google services.
gemini spark is a 24/7 personal agent that works in the background, rolling out to testers. daily brief provides personalized daily digests from connected apps. the gemini app gets a neural expressive redesign with interactive responses and improved live voice. google antigravity 2.0 adds a desktop app, cli, and sdk for multi-agent orchestration. these tools aim to simplify building and managing ai agents for developers and users.
why it matters: these updates make ai more agentic and integrated into daily tasks, from coding to shopping, potentially changing how developers build and users interact with technology.