Google has released a new model with enhanced reasoning performance over 'Gemini 3 Pro.'
Google announced on the 20th that it is launching 'Gemini 3.1 Pro,' its latest artificial intelligence (AI) model optimized for complex high-difficulty tasks and advanced reasoning, and will expand its application across major services.
"Gemini 3.1 Pro, built on the Gemini 3 series, has dramatically advanced core reasoning capabilities," Google emphasized.
In fact, Gemini 3.1 Pro scored 77.1% on the 'ARC-AGI-2' benchmark, which evaluates the ability to solve new logic patterns—double the score of its predecessor, 3 Pro (31.1%). This figure also exceeds competitors OpenAI GPT-5.2 (52.9%) and Anthropic Claude Opus 4.6 (68.8%).
Google believes this model can be used across various domains, including visually implementing complex subjects and integrating data for analysis.
Gemini 3.1 Pro is available starting today, though general consumers on paid subscription plans will receive priority access.