Google has launched Gemini 3.1 Professional, a big replace to its flagship AI mannequin centered on sharper reasoning and higher dealing with of complicated, multi-step duties. The mannequin is rolling out at present throughout the Gemini app, NotebookLM, and developer platforms together with Google AI Studio, Vertex AI, and Antigravity.
Google Gemini 3.1 Professional options: What the improve truly brings
That is Google’s first-ever .1 increment—earlier cycles used .5 because the mid-year refresh. The naming alerts a centered intelligence improve slightly than a broad characteristic growth. At its core, 3.1 Professional brings the superior reasoning engine launched in Gemini 3 Deep Assume final week to a a lot wider viewers.Google describes the mannequin as constructed for conditions the place “a easy reply is not sufficient.” In follow, which means dealing with issues like synthesising massive datasets right into a single view, producing animated SVGs immediately from a textual content immediate, or working by means of layered technical and scientific issues. The animated SVG functionality is a neat instance—outputs are in-built pure code slightly than pixels, that means they keep crisp at any scale with tiny file sizes in comparison with conventional video.
Google Gemini 3.1 Professional vs OpenAI GPT and Anthropic Claude: What the numbers say
The headline determine is a 77.1% rating on ARC-AGI-2, a benchmark that exams summary reasoning on solely new logic patterns. That is greater than double Gemini 3 Professional’s rating—a significant bounce. Elsewhere, 3.1 Professional scores 94.3% on GPQA Diamond (scientific data), 80.6% on SWE-Bench Verified (agentic coding), and 85.9% on BrowseComp (agentic search). On LiveCodeBench Professional’s aggressive coding Elo, it sits at 2887—notably forward of rivals.
Google Gemini 3.1 Professional: How you can entry it proper now
Gemini 3.1 Professional is launching in preview whereas Google validates efficiency, significantly round agentic workflows. Shoppers can entry it by means of the Gemini app and NotebookLM, with increased utilization limits for Google AI Professional and Extremely subscribers. Builders get entry by way of the Gemini API in Google AI Studio, Gemini CLI, Antigravity, Vertex AI, Gemini Enterprise, and Android Studio. Basic availability is predicted quickly.
