Google has a new AI for the most complex tasks. It excels in logical reasoning

A week after the major Gemini 3 Deep Think update, Google comes with another novelty. The Gemini 3.1 Pro model represents an improved version of the core intelligence that also powers the aforementioned Deep Think. The company is starting to distribute it across its products for developers, businesses, and regular users.

Twofold Leap in Reasoning

The main benefit of the new model is its significantly improved logical reasoning capability. In the ARC-AGI-2 benchmark, which measures the ability to solve entirely new logical patterns, Gemini 3.1 Pro achieved a score of 77.1%. This is more than double what the previous 3 Pro model managed. That model was first introduced to the world last November.

Google describes the new model as a tool for tasks where a simple answer is not enough. It is designed to help with analyzing complex topics, synthesizing data from various sources, or creative projects requiring advanced thinking.

Practical Use: Animations Directly from Text

Specific capabilities include generating visually more attractive animated SVGs directly from a text description. Such animations are vector-based, so they remain sharp at any resolution and are a fraction of the size of classic video. They are suitable for websites or presentations, for example.

Where to Find Gemini 3.1 Pro?

The model is currently in preview mode, and Google is gradually expanding its availability. Developers can access it via the Gemini API in AI Studio, Gemini CLI, and Android Studio. Businesses can utilize Vertex AI and Gemini Enterprise. Regular users with a Google AI Pro or Ultra subscription will find it in the Gemini app and in NotebookLM.

Have you tried any of the Gemini 3 series models yet?

Source: Google