On Thursday, Google announced its newest core reasoning model, Gemini 3.1 Pro. As per Google, Gemini 3.1 Pro has achieved performance that is double that of 3 Pro on ARC-AGI-2, a recognized benchmark for assessing a model’s logical reasoning capabilities.

Google first introduced Gemini 3 and 3 Pro in November, and this latest update emphasizes the swift advancement with which AI firms are releasing upgraded models. Gemini 3.1 Pro serves as the latest core model for Gemini and various Google AI applications, such as Gemini 3 Deep Think. Google asserts that it is built to provide more innovative solutions.

A Google blog post describes, “3.1 Pro is designed for scenarios where a straightforward answer falls short, utilizing advanced reasoning to tackle your most challenging problems.” It continues, “This upgraded intelligence can support practical uses — whether you need a straightforward, visual clarification of a complicated subject, a way to amalgamate data into a singular overview, or bring a creative endeavor to fruition.”

Here’s what we currently know about Gemini 3.1 Pro, encompassing its comparison with the latest models from Anthropic and OpenAI, along with how to test it yourself.

How to test Gemini 3.1 Pro

As of today, Google is rolling out Gemini 3.1 Pro within the Gemini App, the Gemini APIA, and in Notebook LM. Free users can experience 3.1 Pro in the Gemini app, while paid users on Google AI Pro and AI Ultra plans will encounter increased usage fees. In Notebook LM, only these paid users will be granted access to 3.1 Pro, at least for the time being. Developers and enterprise users can also utilize the new core model through AI Studio, Antigravity, Vertex AI, Gemini Enterprise, Gemini CLI, and Android Studio.

Gemini 3.1 Pro was already accessible to Mashable editors using Gemini. To experience it yourself, visit Gemini on a desktop or open the Gemini mobile application.