From Text Box to Multimodal Interface
Google is rolling out what it calls its biggest Search revamp in 25 years, driven by the Gemini 3.5 Flash model. The classic single-line text bar now expands into a multimodal search field that accepts long, conversational prompts alongside images, videos, files and even Chrome tabs as inputs. Gemini 3.5 Flash is the default engine behind AI Mode in Search globally, designed to deliver fast, “frontier” performance for both everyday queries and complex tasks. Users can ask follow-up questions directly in Search, with AI Mode handling the conversation-style flow instead of forcing people to rewrite or refine queries manually. Importantly, the traditional experience has not disappeared: users who prefer blue links can still switch to a dedicated Web tab. But the centre of gravity is shifting from simple keyword lookups to a richer, context-aware interface that treats the search bar like a general-purpose command center.
Gemini 3.5 Flash and Agentic Search Capabilities
At the core of this shift is Gemini 3.5 Flash, the first model in the 3.5 series and now the default for the Gemini app and AI Mode in Search. Google says Flash outperforms the earlier Gemini 3.1 Pro on coding and agentic benchmarks, while running up to four times faster than other frontier models. That performance underpins new agentic search capabilities: instead of simply returning results, Search can help complete multi-step tasks. For developers or analysts, Flash can accelerate workflows that previously took days or weeks, particularly when code generation or automation is involved. In Search, these same strengths power dynamic interfaces, follow-up reasoning and real-time updates. Combined with Google’s Antigravity agentic development platform, Gemini 3.5 Flash can assemble interactive tables, graphs and simulations on the fly, effectively designing custom interfaces around a query rather than forcing users to adapt to static pages of links.
Information Agents: Persistent Help Across Multiple Searches
One of the most significant changes for everyday behavior is the arrival of information agents inside Google Search. These AI search agents run continuously in the background, monitoring the web for updates related to a user’s specific task or topic. Instead of searching repeatedly for apartment listings, sneaker drops, stock movements or news about a favourite athlete, you can describe your requirements once and let an agent watch for matching changes. Drawing on blogs, news sites, social posts and fresh data such as finance, shopping and sports, the agent pushes notifications when conditions are met. This turns Search from a pull-based tool into a proactive assistant that spans multiple queries over time. Initially, information agents are slated to launch first for AI Pro and Ultra subscribers, but their design signals a broader shift: search sessions are becoming ongoing projects rather than one-off questions.
Mini Apps and Agentic Coding Inside Search
Beyond answering questions, Google is turning Search into a lightweight app platform. Using agentic coding powered by Gemini 3.5 Flash and the Antigravity system, users can create custom search results and widgets directly from the search bar. Generative UI capabilities allow Search to assemble interactive visuals, dashboards and simulations in real time, such as an astrophysics visualization alongside traditional results. For recurring tasks like planning a wedding, managing a move or building a health routine, Search can generate persistent mini apps—trackers or dashboards you can return to and update over time. A custom fitness tracker, for example, can pull in live reviews, maps, local data and weather to tailor recommendations. These experiences blur the line between browsing and productivity, keeping users inside Search while they organize information, monitor progress and take action, rather than bouncing between multiple external tools and sites.
What This Redesign Means for Everyday Search Behavior
Taken together, multimodal inputs, persistent information agents and mini apps reshape how people will use Google Search. The search bar becomes a multimodal command center where you can speak, type or drop images and files, then refine your intent through conversation. AI search agents continue working after you close the tab, tracking conditions and surfacing changes without repeated manual checks. Agentic interfaces and dashboards turn complex, ongoing tasks into persistent workspaces embedded directly in Search. For many workflows—shopping, research, planning, coding—users may no longer need to leave Google’s environment at all. At the same time, the Web tab preserves traditional link-first search for those who want it. This redesign marks a strategic move: Search is evolving from a directory of the web into an AI-powered operating layer that sits across information, tasks and personal context.