From Keyword Box to Multimodal Command Center
Google’s latest Google Search update replaces the static search box with an expandable, conversational interface powered by Gemini 3.5 Flash. This new default model in AI Mode globally delivers faster, more capable responses, especially for agents and coding, while supporting true multimodal search. Instead of just typing keywords, people can now drop in text, images, videos, files or even Chrome tabs and ask follow-up questions without restarting the session. The field stretches to accommodate long prompts, turning Search into a space for complex instructions rather than one-off queries. Users who prefer traditional links can still tap the Web tab, but the center of gravity is shifting. With Gemini 3.5 Flash, Search starts acting less like a list generator and more like an interactive surface that understands context, keeps track of conversations and orchestrates tasks across formats in real time.
Gemini 3.5 Flash: The Engine Behind AI Search Agents
Gemini 3.5 Flash is the first model in the 3.5 series and now drives AI Mode in Search for everyone. Google says it outperforms Gemini 3.1 Pro on coding and agentic benchmarks while running up to four times faster than other frontier models, making it well suited for AI search agents that must reason continuously. In practice, that performance shows up as smooth follow-up questioning, quicker summaries and more responsive tools that adapt as users refine their tasks. The model also underpins agentic coding inside Search, where it can automatically assemble dashboards, visualizations and widgets on demand. By combining speed, lower compute cost and strong coding capability, Gemini 3.5 Flash search becomes the backbone of a new generation of AI search agents that do more than answer questions—they execute workflows and build interfaces in the background as people explore and iterate.
Information Agents: Always-On Search That Works in the Background
One of the most transformative additions is information agents—AI search agents that keep working after the initial query. Instead of repeatedly searching for the same topic, users can describe what they care about and let Search handle ongoing monitoring. These agents scan blogs, news sites, social posts and fresh data on finance, shopping and sports, then alert users when conditions match their requirements. Apartment hunters, for instance, can list budget, neighborhoods and amenities; the agent watches listings and pings them when a good match appears. Fans can ask to be notified when favorite athletes announce sneaker collaborations or other signature drops. Operating 24/7 in the background, information agents turn Google Search into a persistent research assistant that tracks changes over time. They will roll out first to AI Pro and Ultra subscribers before becoming more broadly available within AI Mode.
Agentic Coding and Mini Apps Turn Search Into a Live Workspace
Beyond monitoring information, Gemini 3.5 Flash enables Search to build things for users through agentic coding. Using Google’s Antigravity platform, Search can generate custom interfaces on the fly—interactive visuals, tables, graphs or simulations that appear alongside results. This generative UI lets people explore topics like astrophysics with dedicated visualizations rather than static text alone. For ongoing projects, Search can go further and craft persistent mini apps, such as wedding planning dashboards, home-move trackers or fitness routines that pull in real-time reviews, live maps and local weather data. These tools live inside Search, so users can return to them, update inputs and track progress over weeks. Instead of forcing people to juggle spreadsheets, notes and browser tabs, the updated experience turns the familiar search field into a workspace where AI designs, codes and maintains tailored tools around long-running tasks.
Personal Intelligence and the Largest Overhaul in Decades
Google frames this shift as its biggest Search revamp in about 25 years, and the changes go beyond the interface. Personal Intelligence in AI Mode is expanding across nearly 200 countries and territories and 98 languages, allowing users to connect services like Gmail and Google Photos for more context-aware help without requiring a premium subscription. That context feeds Gemini-powered agents that can better understand schedules, preferences and past activity while respecting user controls. Meanwhile, agentic booking features in Search can now call home repair, beauty or pet care businesses on a user’s behalf, and an intelligent search box ties AI Mode more tightly to AI Overviews, reducing friction between experiences. With more than 1 billion monthly users already engaging with AI Mode and queries more than doubling every quarter, Google is clearly repositioning Search as an AI platform for persistent, multimodal assistance.