Google has released Gemini 3.1 Pro, an upgraded model building on the Gemini 3 series with significantly improved reasoning capabilities. The model scored 77.1% on the ARC-AGI-2 benchmark, more than double the performance of its predecessor 3 Pro. Available in preview across the Gemini API, Google AI Studio, Vertex AI, the Gemini app, and NotebookLM, the model targets complex tasks including code generation, data synthesis, and creative applications. Google plans to make it generally available soon after validating updates and advancing agentic workflows.

Gemini 3.1 Pro: A smarter model for your most complex tasks - Google Blog