Latest version of Gemini Pro more than doubles the model’s reasoning performance on the ARC-AGI-2 benchmark, the Google Gemini team said. Credit: Shutterstock / Jay Fog Google has released a preview of Gemini 3.1 Pro, described as a smarter model for the most complex problem-solving tasks and a step forward in core reasoning. Announced February 19, Gemini 3.1 Pro is designed for tasks where a simple answer is not enough, taking advanced reasoning and making it useful for the hardest challenges, according to the Google Gemini team. The improved intelligence can help in practical applications such as providing a visual explanation of a complex topic, synthesizing disparate data into a single view, and solving challenges that require deep context and planning. The model is in preview for the developers via the Gemini API in Google AI Studio, Gemini CLI, Google Antigravity, and Android Studio. For enterprises, the model is in Vertex and Gemini Enterprise. Consumers can access Gemini 3.1 Pro via the Gemini app and NotebookLM. Gemini 3.1 Pro follows the Gemini 3.1 release from November 2025. The Gemini team said the core intelligence in Gemini 3.1 Pro also was leveraged in last week’s update to Gemini 3 Deep Think to solve challenges across science, research, and engineering. The team also noted that, on the ARC-AGI-2 benchmark, which evaluates a model’s ability to solve new logic patterns, Gemini 3.1 Pro achieved a verified score of 77.1%, more than double the reasoning performance of Gemini 3 Pro. Artificial IntelligenceSoftware DevelopmentGenerative AI