GitHub Copilot

Market reach

GitHub Copilot is four years old and still the most widely deployed AI coding tool on the market. That staying power comes from one thing: it's everywhere. VS Code, JetBrains, Neovim, Vim, Visual Studio, Xcode, Eclipse, the terminal, GitHub.com, and GitHub Mobile. All from a single subscription. If your team runs across different IDEs, Copilot is the only option that reaches everyone without asking anyone to switch editors.

Completions

The inline completions remain excellent. Near-instant, context-aware, and accurate for mainstream languages and frameworks. Next Edit Suggestions, currently in public preview, predicts where your cursor should jump after an edit based on your recent changes. We used it through a TypeScript refactoring session. It was right about 70% of the time. That's useful enough to change how you move through a file.

The coding agent

The coding agent is more capable than it first looks. Assign a GitHub issue to Copilot and it spins up an ephemeral Actions VM, clones your repo, analyzes the codebase, implements the fix, runs self-review with CodeQL and secret scanning, and opens a draft PR. We gave it a well-scoped bug fix with a clear reproduction path. It came back with a PR that included tests and passed its own review. The safety model — where it can't merge or access main directly — is the right call.

Where it falls short

Where Copilot falls short is complex work. A multi-file refactor spanning ten files produced structural mistakes that Cursor and Claude Code handled correctly. Copilot is strongest on bounded, well-defined tasks: boilerplate generation, unit test scaffolding, API route creation. When the task requires understanding architectural implications across a codebase, the limits show. The 300 premium request cap on Pro is also a real constraint — we burned through it in six days of moderate agent usage.

Enterprise compliance

The most underappreciated part of the Copilot story is enterprise procurement. SOC 2 Type 2, ISO 27001, IP indemnification, and audit logging don't make headlines. They're the reason Copilot ends up in organizations that would otherwise choose a more capable tool. The compliance story clears procurement when other tools don't. For many engineering orgs, that's the only argument that matters.

Our call

Copilot is the right starting point for developers who want AI assistance without disrupting their existing workflow. It's not the strongest tool for complex agentic work — Cursor and Claude Code both beat it there. What it offers instead is the lowest-friction path into AI-assisted coding, a compliance stack that travels through any enterprise procurement process, and a GitHub integration deep enough to feel native rather than plugged in.

Copilot CLI

A terminal-native coding agent that reached GA in February 2026. Plan mode builds structured implementation plans before coding, Autopilot mode runs without approval gates, and /fleet parallelizes work across subagents. Full MCP support and session memory across turns.

Universal IDE Support

Works in VS Code, JetBrains, Neovim, Vim, Visual Studio, Xcode, Eclipse, Azure Data Studio, GitHub.com, and GitHub Mobile. No other AI coding tool runs across this many surfaces from a single subscription.

Agentic Code Review

Rebuilt on a tool-calling architecture in March 2026, the code review agent gathers broader repository context to surface actionable feedback. 71% of reviews produce actionable comments, with 60 million reviews completed to date.

Copilot Spaces

Curated context containers where you organize code, docs, specs, and other content to ground Copilot's responses. Replaces the older knowledge bases with auto-updating, role-based shareable collections that improve suggestion relevance for your specific project.

Next Edit Suggestions

A custom model predicts the next logical edit based on your recent changes and shows an arrow indicator in the gutter. Available in VS Code, JetBrains, Xcode, and Eclipse. Turns inline completion from reactive to proactive editing.