GitHub Copilot

GitHub Copilot is the AI pair programmer that defined a category, launched by GitHub + OpenAI in June 2021 as the first widely-adopted AI coding assistant. Originally based on OpenAI's Codex model and now powered by various frontier models (GPT-4o, Claude Sonnet, custom GitHub models), Copilot is used by 1.8M+ paid subscribers (per GitHub's last public number) plus millions more via GitHub Enterprise. It single-handedly created the 'AI in your IDE' product category that Cursor, Windsurf, and Continue.dev now compete in. What makes Copilot the category default is the distribution + integration combo. It ships natively in VS Code (free for any developer), JetBrains IDEs, Neovim, and Visual Studio. Authentication is GitHub OAuth — instant for anyone with a GitHub account. The Microsoft + GitHub ownership means it's pre-installed for any GitHub Enterprise customer, which covers most large engineering orgs. For 2021-2023, Copilot was the only serious AI coding tool. In 2024-2025, Cursor + Claude Code raised the bar significantly, but Copilot's distribution advantage means it's still the default for most developers. The core feature set: • **Inline code completion** — ghost-text suggestions as you type, accept with Tab. The foundational feature. • **Copilot Chat** — sidebar chat interface for asking questions, generating code, refactoring, explaining • **Multi-file editing** — generate changes across multiple files (Copilot Workspace style) • **Inline chat** — Cmd+I to invoke AI inside specific code blocks for targeted edits • **Code review** — Copilot can review PRs and suggest improvements • **Copilot CLI** — natural language shell command generation (replace memorising flags) • **GitHub.com integration** — explain code, generate code, suggest fixes directly in the GitHub UI • **Multi-model selection** — choose between GPT-4o, Claude Sonnet 4, Gemini 2.5 Pro, custom GitHub models based on task • **Knowledge bases** — index your team's docs/code for context-aware responses • **Test generation** — generate unit tests for selected code • **Documentation generation** — generate docstrings, README files, code comments • **Pull request summaries** — auto-generate PR descriptions from the diff For developers + engineering teams the use cases: • **Daily code completion** — the everyday productivity multiplier; saves 20-40% of typing time for most developers • **Boilerplate generation** — generate repetitive code (CRUD endpoints, test setup, error handling) • **Refactoring** — 'extract this into a function', 'convert to async', 'add error handling' • **Learning new languages/frameworks** — Copilot suggests idiomatic patterns in unfamiliar codebases • **Code review assistance** — auto-summary of PRs, suggested improvements • **Documentation** — generate docstrings + README sections from existing code • **Test coverage** — generate unit tests for legacy code that lacks them • **Enterprise development workflows** — GitHub Enterprise customers get Copilot integrated into their existing workflow The pricing has expanded over time. Copilot Free (2024) provides 2,000 completions/month + 50 chat messages/month for any GitHub user — useful for evaluation and casual use. Copilot Pro at $10/user/month is the standard individual tier (unlimited completions, unlimited chat, model selection). Copilot Pro+ at $39/user/month adds higher model rate limits and premium models. Copilot Business at $19/user/month adds team policies, audit logs, and organisation controls. Copilot Enterprise at $39/user/month adds knowledge bases, custom models, and enterprise SSO. Where Copilot wins clearly: distribution is unmatched (every GitHub user can enable it in 30 seconds), native VS Code integration is the cleanest in the category, the free tier is genuinely useful for evaluation, enterprise sales channel (Microsoft + GitHub) means it's the default for large companies, multi-model selection lets you choose Claude for one task and GPT for another. Where it loses: Cursor + Windsurf have pulled ahead on developer UX for AI-native coding (better multi-file editing, better agentic workflows, better at understanding entire codebases); Claude Code's terminal-first agent workflow is fundamentally different and arguably better for complex tasks; Copilot's response latency lags Cursor for inline completions; the product surface is increasingly fragmented (Chat, Inline Chat, Multi-file, Workspace, Edit mode all overlapping). My take: Copilot is the right call for individual developers who live in VS Code + want the simplest possible AI coding workflow, and for enterprise customers where GitHub Enterprise is already in place. For developers willing to switch IDEs, Cursor genuinely is better for AI-native development. For agentic workflows (build me a feature end-to-end), Claude Code in terminal is fundamentally different. Copilot remains the safe default and the productivity multiplier for traditional VS Code workflows. Pricing at $10/month is reasonable; the value for most professional developers far exceeds that cost.