GitHub Copilot¶

This work is licensed under a Creative Commons Attribution 4.0 International License.
GitHub Copilot is GitHub's AI coding assistant. It began as autocomplete for code and has grown into a family of tools: chat, an agent that edits your project inside the editor, a cloud agent that works on GitHub issues while you do something else, a command-line tool, and automated code review. This page covers what it does as of September 2026, what the plans cost, how students and educators get it free, and how to install it. It is a separate product from the Microsoft 365 Copilot described in the Microsoft Copilot guide.
What GitHub Copilot does now¶
- Inline completions - ghost-text suggestions as you type, from a single line to a whole function. Completions never consume AI credits on any plan.
- Copilot Chat - ask questions about your code, explain errors, and generate tests in a side panel or inline in the editor.
- Agent mode in the IDE - hand Copilot a task and it plans, edits multiple files, runs commands and tests, and iterates until done, asking your approval along the way.
- Copilot cloud agent (formerly "Copilot coding agent") - assign a GitHub issue to Copilot and it researches, plans, and codes in a cloud workspace, then opens a pull request for you to review. Pro plans and above.
- Copilot CLI - the same assistant in your terminal for shell commands, scripts, and repository tasks.
- Copilot code review - automated review comments on pull requests. Pro plans and above; consumes credits.
- Model choice - paid plans let you pick among models from OpenAI, Anthropic (Claude), and Google (Gemini). The Free and Student plans use automatic model selection.
Plans (September 2026)¶
| Plan | Price | Notes |
|---|---|---|
| Free | $0 | 2,000 completions/mo, limited chat and agent-mode usage, automatic model selection |
| Pro | $10/mo | 1,500 AI credits/mo (about $15 of usage); cloud agent and code review; no Claude Opus models |
| Pro+ | $39/mo | 7,000 AI credits/mo (about $70); Claude Opus models |
| Max | $100/mo | 20,000 AI credits/mo (about $200); priority access |
| Business | $19/user/mo | 1,900 credits per user; organisation-managed seats and policies |
| Enterprise | $39/user/mo | 3,900 credits per user; organisation-managed seats and policies |
| Student | $0 | Verified students through GitHub Education; automatic model selection |
| Teachers and open-source maintainers | $0 | Complimentary Copilot Pro through GitHub Education or the maintainer programme |
Billing and sign-ups (September 2026)
- Since June 1, 2026, usage on paid plans is metered in GitHub AI Credits: 1 credit = $0.01, spent per request at each model's rate. Code completions never consume credits. Once a month's credits are used up, further usage is billed per token at model rates.
- New individual sign-ups (Student, Pro, Pro+) were paused on April 20, 2026 and reopened gradually from June 17, 2026, when the Max plan also opened. If a sign-up page still shows a waitlist, check back later; the Free plan stayed open throughout.
- Claude Opus models are available on Pro+ and above only (they were removed from Pro on April 20, 2026).
Full details are on the Copilot plans page and in the subscription plans documentation. To compare with other coding tools, see Choosing the Right AI Platform.
Free access for students and educators¶
- Create a GitHub account if you do not have one, and add your school email address to it (Settings > Emails).
- Go to education.github.com and apply: students choose the Student benefits, faculty and staff choose Teacher.
- Verify your status with your school email address, or upload a dated student ID, transcript, or enrolment letter if your school is not recognised automatically.
- Approval usually takes a few days. Once approved, the GitHub Copilot Student plan (students) or complimentary Copilot Pro (teachers) appears on your account; see Getting free access to Copilot Pro as a student, teacher, or maintainer. Maintainers of popular open-source projects qualify through the same page.
- GitHub Education also unlocks the Student Developer Pack, Codespaces hours, and GitHub Classroom for instructors.
Institutions can join the GitHub Campus Program, which gives the whole school GitHub Enterprise and streamlines student verification; check with your campus IT to find out whether yours participates. UNM users: see CARC.
Install¶
- Install Visual Studio Code (see the VS Code guide for platform-specific steps).
- Press Ctrl+Shift+X (Windows/Linux) or Cmd+Shift+X (macOS) to open the Extensions view.
- Search for GitHub Copilot and click Install on the official GitHub extension. Copilot Chat is installed alongside it.
- Click the Copilot icon in the status bar (or the Accounts icon in the Activity Bar) and Sign in to GitHub; approve the authorisation in your browser.
- Back in VS Code, the status-bar icon turns solid when Copilot is active.
- Open Settings > Plugins in IntelliJ IDEA, PyCharm, WebStorm, or another JetBrains IDE.
- Search the Marketplace for GitHub Copilot and install it; restart the IDE when prompted.
- Choose Tools > GitHub Copilot > Login to GitHub and complete the device-code sign-in in your browser.
- Recent versions of Visual Studio 2022 ship with GitHub Copilot as a built-in component; if it is missing, open the Visual Studio Installer, click Modify, and add the GitHub Copilot component.
- In Visual Studio, click the Copilot badge in the top-right corner and sign in with your GitHub account.
- Install the copilot.vim plugin with your plugin manager (it requires Node.js and Neovim 0.6 or later).
- Run
:Copilot setupand follow the device-code sign-in in your browser. - Run
:Copilot enable; suggestions appear as virtual text and are accepted with Tab.
- Install the standalone Copilot CLI from github.com/github/copilot-cli, or add the older
gh copilotextension to the GitHub CLI withgh extension install github/gh-copilot. - Run
copilot(orgh copilot) in a terminal and sign in with your GitHub account when prompted. - Ask it to explain a command, write a script, or work on the repository in the current directory.
A standalone GitHub Copilot app has been available to all Copilot users since July 2026 (verify). Look for the download link on the Copilot product page and sign in with the same GitHub account.
First use¶
- Completions: open a code file, start typing, or write a comment that describes what you want (
# read a CSV and plot the first two columns). Ghost text appears; press Tab to accept it or keep typing to ignore it. Press Alt+] (Windows/Linux) or Option+] (macOS) to cycle through alternative suggestions. - Chat: press Ctrl+Alt+I (Windows/Linux) or Ctrl+Cmd+I (macOS) in VS Code to open the Chat view, or Ctrl+I / Cmd+I for inline chat on selected code. Ask it to explain a function, write tests, or fix an error message you paste in.
- Agent mode: in the Chat view, switch the mode selector from Ask to Agent, describe a task ("add a command-line flag for the output directory and update the README"), and review each file edit and terminal command before accepting it.
For how Copilot compares with Claude Code, Cline, and other editor assistants, see the VS Code and AI-Powered Development guide; for a broader look at agent-driven programming, see Vibe Coding.