Gemini 3.1 Pro Debuts With Higher Limits and Advanced Reasoning

Google Gemini 3.1 Pro Update

Google has rolled out Gemini 3.1 Pro, three months after Gemini 3 Pro, marking a new step forward in core reasoning. The company describes Gemini 3.1 Pro as a smarter, more capable baseline for complex problem solving and as designed for tasks where a simple answer does not suffice. The update aims to improve reasoning and clarity in everyday applications.

In assessments cited by Google, Gemini 3.1 Pro demonstrates progress in a range of AI benchmarks. The firm notes that the model performs well on ARC-AGI-2, a benchmark that evaluates a model’s ability to solve entirely new logic patterns. The disclosed results show Gemini 3.1 Pro achieved a score of 77.1 percent on ARC-AGI-2, more than double the reasoning performance of Gemini 3 Pro. The figures are presented to illustrate a step forward in core reasoning capabilities.

Google adds that 3.1 Pro is built for tasks where a simple answer is not enough. It uses advanced reasoning to support practical applications such as providing a clear, visual explanation of a complex topic, synthesizing data into a single view, or helping to bring a creative project to life. The company describes 3.1 Pro as a smarter, more capable baseline for complex problem solving and a significant upgrade over earlier models.

Rolling out starting today, 3.1 Pro becomes available to consumers via the Gemini app with higher limits for users on the Google AI Pro and Ultra plans. NotebookLM lists 3.1 Pro as available, but access is limited to Pro and Ultra subscribers. For developers, access is provided in preview through the Gemini API in Google AI Studio, the Gemini CLI, and the agentic development platform Antigravity, along with Android Studio. Enterprises can use 3.1 Pro through Vertex AI and Gemini Enterprise. The company notes that this is a staged rollout intended to validate updates and support further work on ambitious agentic workflows before making a broader general availability launch. The rollouts cover consumer apps, developer tools, and enterprise platforms, signaling alignment across Google’s AI product stack.

In context, 3.1 Pro builds on the Gemini 3 series, representing a step forward in core reasoning. The update forms part of an ongoing push to translate breakthroughs in intelligence into everyday applications. The company characterizes the improvements as enabling clearer, visual explanations of complex topics, easier synthesis of data into a single view, and the ability to bring a creative project to life. In earlier communications, Google described 3.1 Pro as a base for more ambitious problem solving and for tasks where progress in reasoning matters more than a quick answer. The update also signals the company’s intent to push forward in agentic workflows, with further work slated in the near term.

Availability and access details include several lanes for different users. Starting today, the Gemini app expands access for AI Pro and Ultra plan subscribers. NotebookLM provides 3.1 Pro access for Pro and Ultra users. For developers, contact points include the Gemini API via AI Studio, the Gemini CLI, Android Studio, and the Antigravity platform, while enterprise users can reach the model through Vertex AI and Gemini Enterprise. In updates shared with developers and partners, Google emphasizes that the 3.1 Pro upgrade is designed to unlock more robust capabilities in practical settings and to inform future updates through feedback from a broad base of users. The company adds that more improvements are planned as part of ongoing work on agentic workflows and other ambitious tasks.

We are told that the team behind Gemini 3.1 Pro is eager to see what users build and discover with the upgrade, underscoring the plan to publish further progress as more feedback becomes available.

Intelligence gains and benchmarks

Gemini 3.1 Pro is described as a smarter, more capable baseline for complex problem solving that can handle tasks where a simple answer falls short. The model is positioned to deliver advanced reasoning to support a range of practical uses, from visual explanations of complex topics to data synthesis and creative projects. The ARC-AGI-2 benchmark is cited as a key measure, with Gemini 3.1 Pro achieving a verified 77.1 percent, a figure Google says surpasses the prior generation by a wide margin.

• ARC-AGI-2 score: 77.1 percent, cited as a measure of reasoning on novel logic patterns
• Use cases emphasized: clear visual explanations, single-view data synthesis, creative project support
• Comparison baseline: more than double the reasoning performance of Gemini 3 Pro

Rollout and availability

The rollout spans consumer apps, NotebookLM, and developer tools, with previews for API access and enterprise platforms. Availability details include consumer access via the Gemini app with higher limits for AI Pro and Ultra subscribers, NotebookLM access for Pro and Ultra users, and developer access across AI Studio, Gemini CLI, Antigravity, Vertex AI, and Gemini Enterprise. Android Studio is listed among the preview access points, and the rollout also covers enterprise deployments in Vertex AI and Gemini Enterprise.

The update is framed as a continuation of the Gemini 3 series, designed to translate breakthroughs in intelligence into everyday applications for consumers, developers, and enterprises. Google notes that 3.1 Pro is aimed at tasks where reasoning matters more than a quick, surface-level answer and that further progress will follow as the company validates updates through real-world use.

• Consumer access: Gemini app with higher AI Pro and Ultra limits
• NotebookLM access: Pro and Ultra subscribers
• Developer access: Gemini API via AI Studio, Gemini CLI, Antigravity, Android Studio
• Enterprise access: Vertex AI, Gemini Enterprise

We can’t wait to see what users build and discover with the upgrade as more progress is released and more user feedback informs the next steps.