Luisa Crawford
Mar 23, 2026 21:57
Google’s advanced reasoning model Gemini 3.1 Pro now available in GitHub Copilot for JetBrains, Xcode, Eclipse, and Visual Studio users.
GitHub has expanded access to Google’s Gemini 3.1 Pro model across its Copilot AI assistant, bringing the advanced reasoning capabilities to developers working in JetBrains IDEs, Xcode, and Eclipse for the first time.
The integration, announced March 23, makes Gemini 3.1 Pro available in public preview across all Copilot tiers—Enterprise, Business, Pro, and Pro+. Users can access the model through the chat model picker in agent, ask, and edit modes.
What Gemini 3.1 Pro Brings to Copilot
Released by Google DeepMind on February 19, 2026, Gemini 3.1 Pro represents a significant upgrade in reasoning capability. The model scored 77.1% on the ARC-AGI-2 benchmark—double the performance of its predecessor, Gemini 3 Pro.
For developers, the practical benefits include a massive 1 million token context window and expanded output capacity of 65,536 tokens. That context length means the model can process entire codebases, lengthy documentation, or complex multi-file projects in a single prompt.
The model also introduces a three-tier thinking system with low, medium, and high computational modes. Developers can balance reasoning depth against response latency depending on the task—quick syntax questions versus complex architectural decisions.
Enterprise Deployment Requirements
Business and Enterprise administrators need to manually enable Gemini 3.1 Pro through the Copilot settings policy page before their teams can access it. Once activated, the model appears in the standard model picker alongside existing options.
Individual Pro and Pro+ subscribers should see the option automatically, though GitHub notes some users may need to check back if access hasn’t propagated yet.
Broader AI Coding Tool Competition
The expansion puts GitHub in direct competition with other AI coding assistants that have integrated advanced models. By offering Gemini 3.1 Pro alongside its existing model options, GitHub gives developers flexibility to choose based on task requirements rather than locking them into a single AI provider.
The multimodal capabilities of Gemini 3.1 Pro—which can process up to 900 images or one hour of video per prompt—could prove useful for developers working with visual assets, documentation screenshots, or video tutorials alongside code.
GitHub is directing feedback to its Community discussions forum as the feature remains in public preview.
Image source: Shutterstock
