Google’s Gemini 3.7 Flash is rolling out inside GitHub Copilot, giving developers on paid and enterprise plans a new Google-built model option inside the same assistant they already use for code completion and agentic tasks. GitHub confirmed the rollout on August 13, 2026, describing Gemini 3.7 Flash as Google’s latest Flash model and noting that, from the company’s early testing, it improved web and app development and agentic coding workflows compared with its previous version GitHub Changelog.
The model is not a silent background swap. GitHub says Gemini 3.7 Flash also delivers improvements in code quality, final-output presentation, codebase research, and verification during complex coding tasks, according to that same early internal testing GitHub Changelog. Those are vendor-reported gains rather than independent benchmarks, so teams should validate them against their own repositories before assuming a step-change in output quality.
Availability is broad but gated by plan. The model reaches Copilot Pro, Pro+, Max, Business, and Enterprise users and can be selected in the model picker across Visual Studio Code, Visual Studio, Copilot CLI, the GitHub Copilot cloud agent, the Copilot app, JetBrains, Xcode, and Eclipse GitHub Changelog. Rollout is gradual, so some users may not see the option immediately.
Cost follows provider list pricing under usage-based billing, GitHub says, with usage tracked in AI credits GitHub Changelog. GitHub’s pricing documentation states that 1 AI credit equals $0.01 USD and that per-token rates vary by model Models and pricing for GitHub Copilot. The same picker already lists other options, including Microsoft’s MAI-Code-1.1-Flash, which we covered at Microsoft’s MAI-Code-1.1-Flash lands in GitHub Copilot.
Administrators control access for organizations. Copilot Enterprise and Copilot Business administrators must enable the “Gemini 3.7 Flash Preview” policy in Copilot settings before anyone in their org can pick the model GitHub Changelog.
The larger shift is model portability inside the IDE. GitHub documents that supported Copilot models can offer a 1-million-token context window and configurable reasoning levels, features that determine how much code a session can hold and how deeply the model reasons before answering Supported AI models in GitHub Copilot; GitHub’s changelog for this release notes the same expanding model menu GitHub Changelog. Whether Gemini 3.7 Flash exposes those extended settings is worth checking in the picker, since larger context and higher reasoning consume more AI credits.
