Google has launched Gemini 3.1 Pro, the latest version of its Gemini 3 Pro AI model, making it available across developer, enterprise, and consumer platforms from this week.

The release introduces expanded access to Gemini’s updated core intelligence, following the recent upgrade of Gemini 3 Deep Think, which serves as the technical foundation for this new model.

Gemini 3.1 Pro is rolling out in preview for developers via the Gemini API in Google AI Studio, Android Studio, Gemini CLI, and the agentic development environment Google Antigravity.

Enterprise users can utilise the model through Vertex AI and Gemini Enterprise. For consumers, Gemini 3.1 Pro is integrated into the Gemini app and NotebookLM, with enhanced usage limits for Google AI Pro and Ultra plan subscribers.

The model is engineered to strengthen reasoning and problem-solving capabilities in complex scenarios. On the ARC-AGI-2 benchmark, which measures a model’s ability to solve novel logic tasks, Gemini 3.1 Pro achieved a verified score of 77.1%. According to Google, this represents more than double the reasoning performance of Gemini 3 Pro.

In a company blog post, Michael Gerstenhaber, vice president of product management for Vertex AI at Google Cloud, wrote that the launch of Gemini 3.1 Pro marks a step forward in core reasoning based on the Gemini 3 series.

He stated that 3.1 Pro serves as a smarter and more capable baseline for complex problem-solving and that Google is releasing it at scale to support businesses in adapting to an agentic future.

Gerstenhaber wrote: “Gemini 3.1 Pro is designed to solve tougher problems, giving you the reasoning depth your business needs. You can bring disparate data into a single view, visualise complex topics, and solve challenges that require deep context and planning.”

The launch follows last week’s update to Gemini 3 Deep Think, a specialised module developed with researchers and engineers to address scientific and technical challenges where data may be incomplete or solutions undefined. Deep Think is available in the Gemini app for Google AI Ultra subscribers and through the Gemini API for select users in research and enterprise.

Last month, Google introduced Agentic Vision as part of Gemini 3 Flash. Agentic Vision combines visual reasoning with code execution, enabling the model to analyse images through an iterative process rather than as a single static act. By treating vision as an active investigation, the model can plan actions such as zooming in, inspecting, and manipulating images step by step, using code execution to ground its answers in visual evidence.

According to Google, feedback since the November 2025 introduction of Gemini 3 Pro has played a role in driving updates and improvements for this latest release. The company says releasing Gemini 3.1 Pro in preview allows it to validate recent changes while continuing work on agentic workflows before moving to wide general availability.

Consumers with Google AI Pro or Ultra plans can now access higher usage limits within the Gemini app as part of this rollout. NotebookLM also supports the new model exclusively for these subscription tiers.

Earlier this month, Google’s parent firm Alphabet released its financial results for Q4 2025, during which net income reached $34.5bn, an increase of 30% year-over-year from $26.5bn in Q4 2024.

Quarterly revenue rose by 18%, totalling $113.8bn compared to $96.5bn the previous year; operating income increased by 16% to $35.9bn while diluted net income per share climbed to $2.82 from $2.15 over the same period.