Google announced Gemini 3.5 Flash, a new lighter-weight AI model that offers advanced capabilities at about one-third the cost of other frontier models. It will become the default model for the Gemini app and AI mode in Google Search worldwide, CEO Sundar Pichai described the model's performance as “remarkably fast” [1]. The company also previewed Gemini 3.5 Pro, a more powerful model currently used internally, scheduled for broader release next month [1].
Alongside the models, Google launched Gemini Spark, a general-purpose AI agent that integrates deeply with Google apps like Gmail, Calendar, and Docs. This agent can reason across connected apps and act on user instructions. Koray Kavukcuoglu, CTO of Google DeepMind, said, “Before this, I think AI agents were more of an idea in research... this year, they’ll be really in our lives” [2]. Gemini Spark is currently in beta and will be available starting next week to trusted testers and Google AI Ultra subscribers, a paid tier costing about $100 per month [1, 3].
Google also rolled out new AI-powered "information agents" in Search designed to track topics like market trends, flight prices, news, sports, weather, and traffic 24/7. These agents continuously monitor user interests, synthesize data, and provide proactive alerts, effectively replacing Google Alerts with more advanced AI features. The information agents will initially be available to Google AI Pro and Ultra subscribers in the US starting this summer, with plans for an international expansion later [4, 5, 3].
The company unveiled a redesigned Search interface featuring an "intelligent search box" that supports longer, more conversational queries – the biggest change in over 25 years [4, 5, 6]. Google's AI products have seen massive adoption, with AI Overviews in Search reaching over 2.5 billion monthly active users and AI Mode exceeding 1 billion monthly active users [6]. Josh Woodward, VP of Google Gemini, described the interaction style as “like tossing things over your shoulder for the AI to catch and complete” [6].
Google faces some skepticism over widespread adoption of agentic AI given the complexity and multiple entry points, especially as surveys show declining enthusiasm for AI among younger users, including Gen Z [3, 6]. To address safety, Gemini 3.5 Flash incorporates stronger cybersecurity measures to reduce harmful content and minimize false refusals to answer safe queries [1].
Looking ahead, Google expects to release the Gemini 3.5 Pro model more widely by June 2026. Additional AI features such as the Android Halo notification system are planned to ship later this year [1, 3]. The rollout of Gemini Spark to trusted testers starts this week, setting the stage for expanded availability in the months ahead.
