Google has launched a preview of Gemini 3.1 Professional, described as a better mannequin for essentially the most complicated problem-solving duties and a step ahead in core reasoning.
Introduced February 19, Gemini 3.1 Professional is designed for duties the place a easy reply just isn’t sufficient, taking superior reasoning and making it helpful for the toughest challenges, in accordance with the Google Gemini staff. The improved intelligence may help in sensible purposes equivalent to offering a visible rationalization of a fancy matter, synthesizing disparate information right into a single view, and fixing challenges that require deep context and planning. The mannequin is in preview for the builders through the Gemini API in Google AI Studio, Gemini CLI, Google Antigravity, and Android Studio. For enterprises, the mannequin is in Vertex and Gemini Enterprise. Customers can entry Gemini 3.1 Professional through the Gemini app and NotebookLM.
Gemini 3.1 Professional follows the Gemini 3.1 launch from November 2025. The Gemini staff mentioned the core intelligence in Gemini 3.1 Professional additionally was leveraged in final week’s replace to Gemini 3 Deep Assume to resolve challenges throughout science, analysis, and engineering. The staff additionally famous that, on the ARC-AGI-2 benchmark, which evaluates a mannequin’s skill to resolve new logic patterns, Gemini 3.1 Professional achieved a verified rating of 77.1%, greater than double the reasoning efficiency of Gemini 3 Professional.
