Google launched a new product. Your smartphone just became twice as smart

Last week, an update was released Gemini 3 Deep Thinkintended for engineers and researchers. Now the Mountain View giant has made it available more widely new enginewhich powers this scientific tool. Gemini 3.1 Pro he just hit it into the hands of ordinary usersdevelopers and business clients.

Google boasts that Gemini 3.1 Pro is not only cosmetic improvements. This is quite a step forward in what specialists call a field core reasoning, i.e. basic reasoning. The new model is expected to be much smarter and better at solving complex problems.

This improved intelligence can help you with practical applications, whether you’re looking for a clear, visual explanation of a complex topic, a way to synthesize data into a single view, or want to bring a creative project to life

Although these boasts sound like a marketing slogan, the company stands by itself results of popular benchmarks. The result in the ARC-AGI-2 test is particularly impressive, as it tests AI’s ability to solve completely new logical patterns. Gemini 3.1 Pro scored as much as 77.1%which is a more than double performance improvement over its predecessor last November.

Google emphasizes that the new model will work where we expect something more than just a superficial confirmation of facts. Gemini 3.1 Pro can help with visualization of complex issueswhen the user needs a clear and graphic explanation of a difficult topic. The new model is designed to deal with synthesis of huge data sets collected in tables and reports. The announcement shows that Google’s AI will efficiently summarize information in a clear view.

Additionally, Gemini 3.1 Pro can do it from the text description alone generate animated SVG filesready to be placed on the website. The advantage of this option is that the code takes up a fraction of the space of the video, and the graphics remain sharp at any magnification.

The implementation of Gemini 3.1 Pro has already started. It mainly covers ordinary users, a the new model is already in the Gemini application. Free users will also be able to use it, but subscribers to the Google AI Pro and Ultra plans will receive higher query limits. The model will also be available in the NotebookLM application, but here exclusively for holders of paid plans.