Gemini app introduces thinking level controls and third-party app integrations ahead of I/O

You can now give Gemini 'Extended' thinking time to answer your query

You can also set us as a preferred source in Google Search by clicking the button below. The new feature (via 9to5Google) shows up at the bottom of the model picker in the Gemini app. When "Fast" or "Pro" models are selected, users will be able to choose between "Standard" and "Extended" thinking levels. The change seems to be rolling out slowly (typical Google) and is not yet available to everyone. I am unable to see the option on my Gemini app with a Gemini AI Pro subscription. Additionally, Google seems set to roll out new app integrations in Gemini as well. The company's support page now shows instructions for using Canva, Instacart, and OpenTable with Gemini as well. These integrations haven't rolled out yet. However, with the support pages going live, they should be available soon. With Canva, Instacart, and OpenTable, Gemini users will soon be able to design and manage assets, add grocery items to their carts, and make restaurant reservations directly from the app. The integrations will currently be available in English on the Gemini app and website. Gemini is expected to play a prominent role at tomorrow's Google I/O keynote. Expect a new version of Gemini, along with new agentic features on Android devices.

Gemini app rolling out 'Extended' thinking level, new 3rd-party app integrations

Ahead of I/O 2026, Google continues to roll out Gemini app features, with some users now seeing a "Thinking level" option. In the model picker, you'll see a new "Thinking level" menu at the bottom of the sheet. This only appears when selecting Fast (Gemini 3 Flash) or Gemini 3.1 Pro, with Thinking excluded. Your options are "Standard" and "Extended." So far, Thinking level is seeing a limited rollout. This matches how Google AI Studio offers Low, Medium, and High thinking levels. Meanwhile, the Gemini app is preparing to add more integrations with third-party apps. At the moment, third-party apps include @GitHub, @OpenStax, @Spotify, and @WhatsApp. Support documents show that @Canva, @Instacart, and @OpenTable are coming. This is not yet rolled out. Canva You can streamline your creative workflow from generating design ideas in Gemini Apps to managing assets and collaborating on projects in Canva. When you connect Canva to Gemini Apps, you can ask Gemini to help you: You can check the availability of groceries and household items in stores and add items directly to your Instacart shopping cart with Gemini Apps. Your prompt can specify "which store you want to shop at and the version of the product you want to add to your cart," while your "default saved address on Instacart is used to find nearby stores." Example prompts include: OpenTable You can ask Gemini Apps to help you find, book, and manage reservations for restaurants on OpenTable. To do this, Gemini Apps use Reserve with Google to check availability and complete your bookings with OpenTable.

Gemini expands app features with deeper thinking

According to 9to5Google, the Gemini app is already rolling out a new "Thinking level" option for some users, letting them choose between Standard and Extended reasoning modes when using Gemini 3 Flash or Gemini 3.1 Pro. The Extended setting, the outlet notes, is designed for more complex topics that benefit from additional processing time -- a consumer-facing version of the thinking level controls Google AI Studio already offers developers. Gemini is also expanding its roster of third-party apps. GitHub, OpenStax, Spotify, and WhatsApp are already supported, but 9to5Google found support documentation pointing to the additions of Canva, Instacart, and OpenTable. Mashable 101 Fan Fave: Vote for your favorite creator today! The Canva integration would let users generate and edit designs directly through Gemini prompts, while Instacart support would allow users to add ingredients to a shopping cart from a recipe link or a simple list. OpenTable would bring restaurant search, booking, and reservation management into the chat window -- including the ability to hand off a confirmed reservation directly to Google Calendar. None of the three new integrations has rolled out yet, per the report. However, the timing is notable given that Google's I/O developer conference is literally a day away. Perhaps the rollout marks the start of Google transitioning Gemini AI toward an agentic AI rather than a simple chatbot.

Google Gemini's new thinking level lets you dial up the brainpower

With Google I/O 2026 almost here, Google seems unable to stop Gemini leaks from slipping out early. Every other day, something new appears inside the app, and this time it looks like Google is experimenting with giving users more control over how much "thinking" Gemini actually does before responding. According to a report from 9to5Google, some users are now spotting a new "Thinking Level" option inside the Gemini app. The feature reportedly appears within Gemini's existing model picker, where users already choose between options like Fast, Thinking, Pro, or Google AI Plus. Things are getting a little more nuanced Instead of simply choosing which model you want, Google also appears to be testing how deeply that model reasons through a task. The report says the new Thinking Level option currently shows up when selecting Fast (Gemini 3 Flash) or Gemini 3.1 Pro with thinking enabled. For now, the rollout seems extremely limited. If this sounds familiar, that is because Google AI Studio already offers similar controls with Low, Medium, and High reasoning levels. Bringing that flexibility into the regular Gemini app feels like the next obvious step, especially as AI companies increasingly compete on how "thoughtful" or agentic their assistants can feel. Honestly, I feel this could be more useful than it sounds Not every AI request needs maximum reasoning power. Sometimes you just want a quick answer without waiting several extra seconds while the model overanalyzes your grocery list, as if it were preparing a PhD thesis. Giving users control over that balance between speed and deeper reasoning could make Gemini feel much more flexible day to day. Recommended Videos Google also appears to be expanding Gemini's growing ecosystem of third-party app integrations. Right now, Gemini already works with apps and services like GitHub, OpenStax, Spotify, and WhatsApp. However, support documentation reportedly hints that integrations for Canva, Instacart, and OpenTable are also on the way. None of these integrations appears to be live yet, but the timing makes sense. Google I/O is usually where the company shows off Gemini becoming less of a chatbot and more of a proper digital assistant that can actually do things across apps and services. At this point, Gemini's evolution feels less about smarter answers alone and more about turning the app into something that quietly handles parts of your digital life in the background -- ideally without making everything feel unnecessarily complicated.

You Can Choose New 'Thinking Levels' for Gemini

This gives users more control over how much time the Fast or Pro models take to "think" through a request. You need to use a lot of quotation when discussing AI. There's "intelligence," of course, but also "thinking," which these days usually refers to a specific type of model that breaks requests into multiple steps and takes longer to process (or "think through") them. Thinking models are nothing new, but if you use Gemini, you do have new thinking "levels" to choose from when using non-thinking models. Here's how it works. As spotted by 9to5Google, there's a new "Thinking level" option at the bottom of the model selector in Google's Gemini app for iOS and Android. To be clear, this isn't the "Thinking" model listed under "Gemini 3," as this model has been available for some time. Instead, "Thinking level" appears when using either Fast (which uses Gemini 3 Flash), or Gemini 3.1 Pro, when you aren't using Thinking. When "Thinking level" does show up on your end, you'll see two options: "Standard" and "Extended." Google says that Standard is best for most problems, while Extended offers more time for "complex topics." It seems that Google is giving users more processing options when using non-thinking models in Gemini. If you're using the Fast option, for example, you might not get the answer (or the quality of answer) you're looking for, since the model is designed to produce a response as quickly as possible. Therefore, you may want to dive into the Thinking level menu to switch to Extended: It may take a bit longer than usual, but you may boost your chance of getting the response you're expecting -- all without having to use the actual Thinking model, which may take too long altogether. Per 9to5Google's coverage, this doesn't appear to be exclusive to Google AI Plus subscribers. In the outlet's screenshots showing the feature, there's still an option to upgrade to the subscription plan, so free users should see "Thinking level" rolling out on their end. (As of this article, I don't see it yet.) These changes come on the eve of Google I/O, where Google will undoubtedly announce many new Gemini features and AI updates in general. Last week, the company hosted The Android Show: I/O Edition, which ran through a number of Android announcements ahead of I/O. This was the second year in a row Google decided to break out Android news into its own segment, likely because it wants to dedicate so much of I/O to AI.

Google Gemini may soon think deeper, book restaurants and shop groceries for you

The new Gemini features are currently said to be available only to a limited number of testers. Right before IO 2026, Google is preparing new upgrades for the Gemini app including deeper third party integrations and new Thinking level option for AI responses. Some users have reportedly started seeing these additions in limited rollout phases across the Gemini. While other things are not announced, the Thinking level setting inside the Gemini model picker is among the most notable upgrades. The option reportedly appears when users select the Gemini 3 Flash or Gemini 3.1 Pro models. The users can opt for Standard or Extended reasoning modes that can allow them to control how deeply the AI process responds. The feature is being rolled out to a small number of users and appears to be similar to the reasoning controls that are already available in Google AI Studio. Also read: Google may no longer give 15GB free storage to everyone: Here is the catch Aside from model controls, Google is also working to improve Gemini's integration with third-party services. Currently, Gemini supports applications such as GitHub, Spotify, WhatsApp, and OpenStax. However, support pages now indicate that integrations with Canva, Instacart, and OpenTable may be coming soon, though these are not yet publicly available. The Canva integration is said to enable users to create and manage designs directly from Gemini. As per the reports, the users can use prompts to create invitation cards, presentations and marketing assets. It may also move files between folders, respond to comments and edit generated images with Canva. On the other hand, Instacart integration can change Gemini into a shopping assistant. The reports state that it may become capable of checking grocery availability at nearby stores. Lastly, OpenTable support can help in restaurant discovery and reservation management. Gemini may enable users to search for restaurants, check table availability, book reservations, modify bookings and cancel reservations using natural language prompts. So far, the Thinking Level setting and several third party integrations appear to be in limited testing. Google has not officially announced a wider rollout timeline yet.