Three months after unveiling Gemini 3 Pro on November 19, Google is back with another upgrade. The company has introduced Gemini 3.1 Pro, calling it a meaningful jump in reasoning and problem-solving. This time, Google is focusing less on flashy claims and more on measurable gains.
According to Google, Gemini 3.1 Pro delivers stronger core reasoning. The company describes it as a smarter and more capable baseline for complex tasks. In simple terms, it is built to handle harder questions with better logic and clearer thinking.
Those claims show up in benchmark results. Google shared comparisons between Gemini 3.1 Pro, Gemini 3 Pro, and rival models. In several AI tests, the new version performs ahead of competitors. Most notably, it stands out in ARC-AGI-2. Google describes ARC-AGI-2 as a benchmark that tests how well a model can solve entirely new logic patterns. On that test, Gemini 3.1 Pro ranks at the top.
However, Google is not positioning this model for quick answers. Instead, it says Gemini 3.1 Pro is designed for situations where a simple response will not do. The model uses advanced reasoning to tackle difficult challenges. For example, it can help users understand a complex topic through clear visual explanations. It can also combine scattered data into one organized view. In addition, it supports creative projects that require structured thinking and deeper insight.
In practical use, the improvements could matter more than benchmark wins. Stronger reasoning often translates into better research summaries, clearer breakdowns, and smarter workflows. That is where Google appears to be aiming this release.
Starting today, Gemini 3.1 Pro is rolling out to all users through the Gemini app. That said, Google AI Pro and Ultra subscribers will receive higher usage limits. The model is also available inside NotebookLM, but access there is limited to Google AI Pro and Ultra subscribers.