Google has introduced Gemini 3.1 Pro, an updated AI model that has set new benchmark records.
This is one of the most powerful solutions on the market.
“The new version has advanced in fundamental reasoning. It is a smarter and more capable neural network for solving complex tasks,” the announcement states.
In the ARC-AGI-2 test, which evaluates the model’s ability to solve new logical patterns, it achieved a score of 77.1%. This result more than doubled the performance of the previous Gemini 3 Pro version.
Gemini 3.1 Pro is available to:
- developers — in preview mode via the Gemini API in Google AI Studio, Gemini CLI, Google Antigravity, and Android Studio;
- enterprise clients — on the Vertex AI and Gemini Enterprise platforms;
- the general public — through the Gemini app and NotebookLM.
“Gemini 3.1 Pro is suitable for tasks where a simple answer is insufficient — it employs advanced reasoning to tackle the toughest challenges,” the developers noted.
They added that the enhanced intelligence aids in practical scenarios, whether it is providing a clear and understandable explanation of a complex topic, integrating data into a cohesive picture, or bringing a creative project to life.
Back in February, Google updated the reasoning mode of Gemini 3 Deep Think. The tool is positioned as a solution for complex tasks in science and engineering.