The Bottom Line: An AI coding agent built into GitHub, covering the full development workflow—exploring codebases, planning, automating tasks, reviewing changes, and resolving merge conflicts—rather than just autocomplete. It works across VS Code, Visual Studio, JetBrains IDEs, Xcode, Vim, Neovim, Zed, Eclipse, the terminal/CLI, a desktop app, the web, mobile, Slack, and Teams, with sessions that continue across surfaces. Pricing was restructured in 2026 around GitHub AI Credits (1 credit = $0.01): code completions and next edit suggestions stay unlimited on paid plans, while chat and agent sessions consume credits based on model and task complexity. Pro ($10/mo) includes $15 monthly credits and access to third-party agents (Claude Code, Codex); Pro+ ($39/mo) adds premium models like Opus and $70 credits; Max ($100/mo) adds priority access and $200 credits. Privacy note: individual subscribers' interaction data may be used to train AI models, with an opt-out available in settings; Business/Enterprise data is not used this way.
Pricing model: Freemium — a permanent free tier with 2,000 completions per month and access to entry-level models (Haiku 4.5, GPT-5 mini, Copilot Fusion Standard), no credit card required . Pro is $10/user/month with unlimited code completions, access to Copilot in the cloud and code review, third-party agents (Claude Code, Codex), model selection, and $15 monthly AI credits. Pro+ is $39/user/month with premium models including Opus, audit logs, 4×+ included usage versus Pro, and $70 monthly credits. Max is $100/user/month with priority access, 2.9×+ included usage versus Pro+, and $200 monthly credits . Credits draw down at $0.01 each; when the monthly allowance runs out, users can wait for the next cycle, set a paid dollar budget for additional usage, or switch to a cheaper model . Business and Enterprise add admin-controlled usage limits .
Platforms: Web, macOS, Windows, Linux (via IDE extensions, CLI, and desktop app), iOS, Android. Official availability spans VS Code, Visual Studio, JetBrains IDEs, Xcode, Vim, Neovim, Zed, Eclipse, the terminal/CLI, the Copilot desktop app, the web, mobile, Slack, and Teams, with cross-surface session continuity .
Pros & Cons
Pros: Deep integration with the GitHub workflow is the core differentiator—issues, pull requests, code review, and merge conflicts are all handled in the same place, rather than exporting code to a separate tool. GitHub's own documentation describes a change starting as an issue, being picked up by an agent, returning as a pull request, getting reviewed for quality and security, and being merged, with the user and Copilot working in the same place with the same context . Code completions and next edit suggestions are unlimited on paid plans and do not consume credits, so routine coding does not burn the allowance . Third-party agents (Claude Code, Codex) are included from Pro upward, giving model choice without separate subscriptions . Multi-surface support with cross-surface session continuity is broader than most competitors. The free tier with 2,000 completions per month supports evaluation without payment .
Cons: The 2026 credit model adds cost uncertainty—while completions are unlimited, agent and chat sessions consume credits at rates that depend on model and task complexity, and hitting the monthly allowance requires waiting, setting a paid budget, or downgrading to a cheaper model . Individual subscriber data may be used for model training by default. GitHub's own privacy documentation states that for individual subscribers only, GitHub may use Copilot interaction data—including prompts, suggestions, and code snippets—to train and improve AI models, with an opt-out available in settings . This is a meaningful privacy difference from Business/Enterprise and should be surfaced explicitly. Premium models like Opus require Pro+ at $39/mo, a significant jump from Pro. The product is GitHub-centric—teams on GitLab or Bitbucket get less value from the integrated workflow. Independent large-scale third-party review volume beyond GitHub's own documentation is limited.
Who is it for?
Ideal for: Developers already working in the GitHub ecosystem who want AI assistance without leaving issues, PRs, and code review; teams that want a single AI assistant across multiple IDEs and surfaces with session continuity; individual developers who want a functional free tier or a $10 entry point with third-party agent access; organizations on Business/Enterprise that need data not used for model training and admin-controlled usage limits.