Gemini 3.1 Pro Now in Public Preview in GitHub Copilot

Gemini 3.1 Pro Enters GitHub Copilot Public Preview

Google's Gemini 3.1 Pro has arrived in GitHub Copilot as a public preview model. The release follows Google's broader launch of Gemini 3.1 Pro on February 19, 2026, and represents a swift integration of Google's most recent model generation into Copilot's multi-model ecosystem.

Performance Characteristics

In early testing, Gemini 3.1 Pro distinguishes itself through efficient agentic task execution. GitHub describes its strength as "effective and efficient edit-then-test loops with high tool precision," meaning the model achieves strong results while requiring fewer tool interactions per benchmark compared to its predecessor. Google has reported a score of 77.1% on ARC-AGI-2, more than double Gemini 3 Pro's 31.1%, along with 80.6% on SWE-Bench Verified.

Where It's Available

Gemini 3.1 Pro is accessible via the model picker on the following platforms:

• Visual Studio Code (chat, ask, edit, and agent modes)
• Visual Studio (agent and ask modes)
• github.com interface
• GitHub Mobile (iOS and Android)

The rollout is gradual, so some users may not see the model immediately.

Eligibility and Administrative Setup

The model is available to Copilot Pro, Pro+, Business, and Enterprise subscribers. For Business and Enterprise plans, administrators must explicitly enable the Gemini 3.1 Pro policy through Copilot settings.

What This Means for the Copilot Model Landscape

Gemini 3.1 Pro's arrival continues GitHub's strategy of offering access to the latest frontier models across major AI providers — Anthropic, OpenAI, and Google — within a single subscription.