GPT-5.4 mini Now Available in GitHub Copilot

GitHub is rolling out OpenAI's GPT-5.4 mini as a new model option for GitHub Copilot users. This lightweight version of the GPT-5.4 agentic coding model is designed for speed and efficiency, delivering the fastest time-to-first-token response and enhanced capabilities for codebase exploration. Early testing shows it performs better than previous mini models, particularly when using grep-style tools for code search and analysis.

Availability and Pricing

GPT-5.4 mini is available to all GitHub Copilot subscribers:

• Copilot Pro, Pro+, Business, and Enterprise users can access the model
• The model carries a 0.33x premium request multiplier on standard pricing (note: pricing is tentative and subject to change)
• Access is available immediately across the full Copilot ecosystem

Where to Use It

The model can be selected from the model picker in:

• IDEs: Visual Studio Code, Visual Studio, JetBrains, Xcode, Eclipse
• Web & Mobile: github.com, GitHub Mobile (iOS/Android), and GitHub CLI
• All chat modes: chat, ask, edit, and agent modes (depending on IDE support)

For optimal performance, GitHub recommends upgrading to the latest version of each IDE, as newer versions provide better prompt handling and model parameter tuning.

Enabling Access

Enterprise and Business administrators must enable the GPT-5.4 mini policy in Copilot settings to make it available to their teams. Bring Your Own Key users can enable the model by selecting "Manage Models" from the picker and providing their OpenAI API key.

For complete details on available models, see the GitHub Copilot documentation.