Google gates Gemini to flagships

Google is drawing a line between its newest Android AI features and the broader Android device base. On May 12, the company said Gemini Intelligence would begin rolling out “in waves” on “the latest Samsung Galaxy and Google Pixel phones” this summer, ahead of wider expansion to watches, cars, glasses and laptops later in 2026. Google did not publish a full compatibility list in its announcement, but it said the features were built for its “most advanced devices.” At the same time, the Gemini app is adding a new “Thinking level” control and more app connections, according to 9to5Google. ### Which phones did Google actually name? Google named two devices in its own product post: the Galaxy S26 and Pixel 10. The company said it had spent months fine-tuning multi-step automation capabilities on those phones in food and rideshare apps to make interactions feel seamless. Google used broader language elsewhere. In its Android Developers blog, the company said Gemini Intelligence brings new features to its “most advanced Android devices” and that the rollout will start with the latest Samsung Galaxy and Google Pixel phones this summer. ### What is Google putting behind that gate? Google said Gemini Intelligence is designed to automate multi-step tasks across apps, summarize web pages in Chrome, help fill out forms, rewrite spoken thoughts into polished messages through a feature called Rambler, and generate custom widgets from natural-language prompts. (blog.google) The company said those features combine premium hardware with software integration and would arrive in stages. (developer.android.com) Matthew McCullough, Google’s vice president of product management for Android Developer, said the system is meant to let Android devices handle more of the “heavy lifting” of anticipating user needs, while apps expose actions through automation and AppFunctions. He said Google has begun testing early-stage APIs in private preview with apps including KakaoTalk. (blog.google) ### What is changing inside the Gemini app right now? 9to5Google reported on May 17 that some users are seeing a new “Thinking level” option in the Gemini app, including an “Extended” mode, ahead of Google I/O 2026. The publication also reported that Google is adding new third-party app integrations. Google’s own support pages already show Gemini connected-app features reaching beyond first-party services. (developer.android.com) The company says Gemini can create events in Google Calendar or third-party apps, play music from Spotify, search Google Photos or Samsung Gallery, and make calls or send messages with default apps or WhatsApp on Android. (9to5google.com) ### Why does the device split matter for Android apps? Google’s developer guidance says Gemini Intelligence will arrive in waves and across multiple form factors, while app developers can support it either through “no-code change” automation or through AppFunctions APIs for more direct control. That means developers may face different capability levels depending on the device a user has, even within Android. (support.google.com) Google has not publicly framed that as a limitation, but its own rollout language points to staggered availability rather than a platform-wide launch. The company said broader access across watches, cars, glasses and laptops would come later this year, after the initial phone release. ### What has Google said about timing? Google said on May 12 that Gemini Intelligence would start on select Samsung Galaxy and Google Pixel phones “this summer.” The company did not give a specific launch day or publish pricing tied to the feature in the announcement. (developer.android.com) Google I/O 2026 begins on May 19 at Shoreline Amphitheatre in Mountain View, California, and runs through May 20 online and in person. (blog.google) Google has said the event will include updates across Gemini and Android, making it the next scheduled venue for fuller details on device support and rollout. (blog.google)