Back to List
TechnologyAIInnovationMachine Learning

Google Unveils Gemini 3.1 Pro: A 'Deep Think Mini' with Dynamic Adjustable Reasoning for Enterprise AI and Developers

Google has released Gemini 3.1 Pro, an incremental update to its frontier AI model, introducing a significant innovation: a three-tier adjustable thinking system. This feature allows the model to dynamically scale its reasoning effort, effectively transforming it into a lightweight version of Google's specialized Deep Think reasoning system. This marks a shift in Google's release strategy towards more frequent, incremental upgrades. The new low, medium, and high thinking levels provide enterprise AI teams, developers, and IT leaders with a single model capable of handling tasks from quick routine queries to multi-minute deep reasoning for complex problems. Gemini 3.1 Pro is now rolling out in preview across various Google platforms, including the Gemini API, Google AI Studio, Vertex AI, and the consumer Gemini app, offering fine-grained control over computational effort per response.

VentureBeat

For the past three months, Google's Gemini 3 Pro has maintained its position as one of the most capable frontier models available. However, in the rapidly evolving AI landscape, three months is a significant period, and competitors have continued to advance. Earlier today, Google launched Gemini 3.1 Pro, an update that introduces a crucial innovation to the company's powerful workhorse model: three levels of adjustable thinking. This effectively transforms it into a lightweight version of Google's specialized Deep Think reasoning system.

This release signifies the first time Google has issued a "point one" update to a Gemini model, indicating a change in the company's release strategy from periodic full-version launches to more frequent incremental upgrades. More importantly for enterprise AI teams evaluating their model stack, 3.1 Pro's new three-tier thinking system—low, medium, and high—provides developers and IT leaders with a single model that can dynamically scale its reasoning effort. This allows for quick responses to routine queries and extends to multi-minute deep reasoning sessions for complex problems.

The model is currently rolling out in preview across various platforms. These include the Gemini API via Google AI Studio, Gemini CLI, Google's agentic development platform Antigravity, Vertex AI, Gemini Enterprise, Android Studio, the consumer Gemini app, and NotebookLM.

The most impactful feature in Gemini 3.1 Pro is not a single benchmark number, but rather the introduction of a three-tier thinking level system. This system gives users fine-grained control over the amount of computational effort the model invests in each response. Gemini 3 Pro previously offered only two thinking modes: low and high. The new 3.1 Pro adds a medium setting, which is similar to the previous high setting. Crucially, it also overhauls the meaning of "high." When set to high, 3.1 Pro functions as a "mini version of Gemini Deep Think," which is the company's specialized reasoning model that received an update just last week.

Related News

Technology

Qwen-Agent: A New Framework and Application Built on Qwen>=3.0, Featuring Function Calling, MCP, Code Interpreter, RAG, and Chrome Extension Support

Qwen-Agent is a newly introduced framework and application developed by QwenLM, built upon the Qwen>=3.0 foundation. This versatile agent supports a range of advanced functionalities, including function calling, MCP (Multi-modal Co-reasoning and Planning), a code interpreter, Retrieval-Augmented Generation (RAG), and integration with Chrome extensions. The project, published on March 8, 2026, is available on GitHub Trending, indicating its relevance and potential interest within the developer community.

Technology

AI-Driven Hedge Fund Concept: Exploring Artificial Intelligence for Trading Decisions in an Educational Proof-of-Concept Project

A new project, 'AI Hedge Fund,' has emerged as a proof-of-concept for an AI-driven hedge fund team. Developed by virattt and trending on GitHub since March 8, 2026, this initiative aims to explore the application of artificial intelligence in making trading decisions. The project is explicitly stated to be for educational purposes only, focusing on demonstrating the potential of AI in financial markets without implying real-world investment advice or functionality.

Technology

Agency-Agents: A Complete AI Agency with Specialized Experts, from Frontend Wizards to Reddit Community Ninjas

Agency-Agents, a project trending on GitHub, introduces a comprehensive AI agency where each agent acts as a specialized expert. These AI agents are designed with distinct personalities, processes, and proven deliverables, covering roles from frontend wizardry to Reddit community management, and from idea generation to reality checking. The initiative aims to provide a full suite of AI experts readily available.