Executive Summary
Google has announced the immediate availability of Gemini 3.1 Pro Preview through its AI Gateway service. This new model focuses on quality improvements for specialized tasks, particularly in software engineering, agentic workflows, finance, and spreadsheet applications. The release emphasizes greater efficiency by reducing token consumption and introduces a new "medium thinking level" to give developers finer control over the balance between cost, speed, and performance.
Key Takeaways
* Model Availability: Gemini 3.1 Pro Preview is now accessible for use via the AI Gateway.
* Targeted Improvements: The model is optimized for software engineering, agentic workflows, and real-world tasks in finance and spreadsheets.
* Increased Efficiency: It features more efficient thinking processes, designed to reduce token consumption while maintaining high performance.
* New "Thinking Level" Control: A "medium thinking level" parameter has been introduced, allowing users to make trade-offs between cost, performance, and speed.
* Implementation: Developers can access the model by setting the model parameter to `google/gemini-3.1-pro-preview` in their AI SDK calls.
Strategic Importance
This release signals Google's strategy to refine its AI models for specific, high-value enterprise use cases while addressing customer concerns around operational cost and efficiency.