Google released Gemini 3 Pro on November 19, and now, exactly three months later, it is unveiling Gemini 3.1 Pro. According to Google, this model signifies a notable advancement in core reasoning capabilities, serving as "a smarter, more capable baseline for complex problem-solving."
The enhancements in Gemini 3.1 Pro are evident through its performance on various AI benchmarks. The comparison of Gemini 3.1 Pro with other recent models from Google’s competitors, as well as Gemini 3 Pro, is illustrated in the image below.
Google emphasizes the model's performance in the ARC-AGI-2 benchmark, which assesses a model's capability to tackle entirely new logic patterns. Gemini 3.1 Pro achieved the highest score in this category.
This new model is tailored for tasks where straightforward answers are insufficient, employing "advanced reasoning" to address more complex challenges. Gemini 3.1 Pro's enhanced intelligence can assist in practical scenarios, whether that means offering a clear, visual explanation of a complicated topic, synthesizing data into a cohesive view, or bringing a creative project to fruition.
Gemini 3.1 Pro is being rolled out today via the Gemini app, with higher limits for Google AI Pro and Ultra subscribers. This model is also available in NotebookLM but exclusively for subscribers of the Google AI Pro and Ultra plans.
Source
