Google rolls out Gemini 3.1 Pro, a smarter model for the most complex tasks
Gemini 3.1 Pro advances core reasoning and complex problem-solving, tops ARC-AGI-2 benchmark, and rolls out with tiered limits and NotebookLM availability.
Google announces Gemini 3.1 Pro, says it's better at complex problem-solving
Google released Gemini 3.1 Pro, improving problem-solving and reasoning with higher benchmark scores, notably ARC-AGI-2 (77.1%) and Humanity's Last Exam (44.4%).
