Gemini 3.5 Flash rolled into Google Search and other flagship AI products

Google said on May 19 that Gemini 3.5 Flash is now the default model in the Gemini app and AI Mode in Search globally, extending the company’s newest lightweight AI model into two of its highest-traffic consumer products. The change was announced alongside Google I/O 2026 and means many users may encounter the model without choosing it manually or paying for a premium tier. Google said the rollout covers “billions of people globally” through Search and the Gemini app. ### Where is Gemini 3.5 Flash showing up right now? Google’s May 19 blog post said Gemini 3.5 Flash is now the default model for the Gemini app and AI Mode in Search globally. That places the model inside Google’s standalone AI assistant and inside the company’s AI-driven Search experience, rather than limiting it to developer tools or paid subscribers. (blog.google) CNBC reported after a pre-I/O briefing that Google described 3.5 Flash as the default model in those products worldwide. Google’s broader I/O roundup also listed Gemini 3.5 Flash as one of the main launches from the event. ### Does this mean free users are already using it? Google said 3.5 Flash is available “for everyone via the Gemini app and AI Mode in Google Search,” language that indicates access is not confined to Google AI paid plans. (blog.google) Lifehacker separately reported on May 20 that users could already try several new I/O features for free and said many people may have already used Gemini 3.5 Flash in Search or the app over the prior 24 hours. (cnbc.com) Google’s help documentation also says some Gemini app features can be used without signing in, depending on device type and location, though access varies by feature and market. The company’s subscriber help page separately says paid plans provide expanded access to features and models rather than exclusive access to Gemini apps as a whole. (blog.google) ### What is Google saying is different about 3.5 Flash? Google said 3.5 Flash combines “frontier intelligence with action,” framing it as a faster model built for agentic tasks rather than only chat responses. In its model announcement, the company said the system is powering new features meant to bring “frontier-level intelligence” into everyday products. (support.google.com) Sundar Pichai told reporters, according to CNBC, that Gemini 3.5 Flash is “remarkably fast.” Google also said in its materials that users “no longer have to trade quality for latency,” and its developer post said the model runs four times faster than other frontier models while outperforming Gemini 3.1 Pro across almost all benchmarks. That performance comparison came from Google’s own presentation materials. (blog.google) ### If it is free to reach, what still sits behind limits? Google’s support page says Gemini Apps now use compute-based limits that refresh every five hours until a weekly limit is reached. The company said those limits depend on prompt complexity, the models and features used, and the length of a chat. (cnbc.com) The same help page says users can upgrade to a Google AI plan for expanded access. That means the current shift is less about making Gemini unrestricted and more about putting the newest default model in front of a broader base of users while keeping paid tiers for higher usage and added features. That characterization is an inference from Google’s rollout and support pages. (support.google.com) ### What comes next from Google’s model lineup? Google said on May 19 that Gemini 3.5 Pro is already being used internally and will roll out next month. The company’s I/O collection page also said Gemini Omni and Gemini 3.5 were part of the latest model launches announced this week. (blog.google) Google’s next public milestones are likely to come through its Gemini product updates page, model posts on The Keyword, and support documentation for Gemini Apps. As of May 21, the consumer-facing change already in place is the default switch to Gemini 3.5 Flash in Search AI Mode and the Gemini app. (blog.google 1) (blog.google 2)