Google introduced Gemini 3.1 Pro three months after the release of the previous version
Google introduced Gemini 3.1 Pro — a new version of its artificial intelligence model. The previous Gemini 3 Pro was released by the company on November 19, and three months later, the update was announced. Google claims that the model has undergone changes in basic reasoning mechanisms and is geared towards more complex usage scenarios.
What is known
According to the company, Gemini 3.1 Pro showed results in a number of industry benchmarks. Specifically, in the ARC-AGI-2 test, which checks the model's ability to work with new logical patterns, it demonstrated the highest score among the compared models. Google also published a comparison with Gemini 3 Pro and solutions from other developers.
Comparative results table of AI models. Illustration: Google
The company notes that the model is intended for tasks requiring extended analysis and multi-level information processing. Possible usage scenarios include data summarization, creation of structured reviews, preparation of visual explanations of complex topics, and support for creative projects.
Gemini 3.1 Pro began rolling out from February 19 in the Gemini app for all users. Google AI Pro and Ultra subscribers receive higher access limits. Additionally, the model is integrated into NotebookLM, where it is available for users of paid Google AI plans.
Source: Google