Google Gemini 3.1 Pro: A Deep Dive into the Latest AI Model
Google has released Gemini 3.1 Pro, a significant update to its workhorse AI model, introducing a three-tier reasoning system that mirrors the capabilities of its specialized Deep Think model. This release marks a shift towards more frequent, incremental upgrades and offers enterprise AI teams a scalable solution for varying task complexities.
The ‘Deep Think Mini’ Effect: Adjustable Reasoning on Demand
Gemini 3 Pro previously offered only two thinking modes: low and high. Gemini 3.1 Pro adds a medium setting and overhauls the “high” setting, making it function as a “mini version of Gemini Deep Think.”1 This allows organizations to use a single model endpoint and adjust reasoning depth based on the task, optimizing for both speed and accuracy.
Benchmark Performance: Significant Improvements in Reasoning and Agentic Capabilities
Google’s benchmarks demonstrate substantial improvements, particularly in reasoning and agentic capabilities. On ARC-AGI-2, a benchmark evaluating abstract reasoning, Gemini 3.1 Pro scored 77.1%, more than double the 31.1% achieved by Gemini 3 Pro and surpassing competitors like Anthropic’s Sonnet 4.6 (58.3%) and Opus 4.6 (68.8%), as well as OpenAI’s GPT-5.2 (52.9%).3
Further gains were observed on Humanity’s Last Exam (44.4% without tools, up from 37.5% for Gemini 3 Pro) and GPQA Diamond (94.3%).3 The improvements are particularly relevant for enterprise AI teams in agentic benchmarks.
- Terminal-Bench 2.0: 68.5% (vs. 56.9% for Gemini 3 Pro) 3
- MCP Atlas: 69.2% (vs. 54.1% for Gemini 3 Pro) 3
- BrowseComp: 85.9% (vs. 59.2% for Gemini 3 Pro) 3
Why a ‘0.1’ Release?
Google’s decision to designate this update as 3.1, rather than a full 3 Pro release, signals that the improvements are substantial enough to warrant a version increment while setting expectations for ongoing evolution. The update builds on lessons from the Gemini Deep Think series and leverages reinforcement learning, particularly in areas like reasoning, coding and agentic evaluations.4
Availability
Gemini 3.1 Pro is currently available in preview through the Gemini API via Google AI Studio, Gemini CLI, Google Antigravity, and Android Studio for developers. Enterprise customers can access it through Vertex AI and Gemini Enterprise. Consumers on Google AI Pro and Ultra plans can access it through the Gemini app and NotebookLM.
Key Takeaways
- Gemini 3.1 Pro introduces a three-tier reasoning system, offering adjustable computational effort.
- The model demonstrates significant performance gains across various benchmarks, particularly in reasoning and agentic capabilities.
- The ‘0.1’ release signifies Google’s shift towards more frequent, incremental updates.
- Gemini 3.1 Pro is available in preview for developers and enterprise customers.
Worth a look