GILD verdict
Choose Claude Code when engineers want Anthropic's agent as their main tool, steered from the terminal, an IDE or the desktop app.
Choose Copilot when GitHub is your system of record and you want one policy and one bill across model providers. Outcome studies split. In Microsoft's early 2026 rollout, engineers merged 24.9% more pull requests in weeks they used Copilot CLI and 11.4% more in weeks they used Claude Code. Microsoft, which ran that study, owns GitHub. In a study of popular Python repositories with data to November 2025, a statistical model estimates that 84.3% of Claude Code's agent pull requests merge against 59.6% of Copilot's. Copilot can also run Anthropic's Claude agent on GitHub, a feature not yet generally available, so test whether it covers your delegation before you pay for both. A Claude Team seat lists at US$25 a month, US$6 more than Copilot Business at US$19, and US$1 more when billed annually at US$20.
Decide where delegated work should start, in GitHub or at an engineer's terminal. Copilot is bought as GitHub seats with pooled AI credits and models from several providers, and Claude Code is Anthropic's agent, bought with Claude plans or API usage. Both run agents in the terminal, IDEs, a desktop app and the cloud, and both review pull requests.
Choosing between Claude Code and GitHub Copilot
Choose Claude Code if
- Engineers want Anthropic's agent as their main tool, steered from the terminal, an IDE, or the Claude desktop app.
- You want Claude chat and Claude Code on one seat, drawing from one usage allowance.
- You want permission modes your organization can set, from auto mode, where a classifier blocks unsafe actions, to a manual mode that starts read-only, with CLAUDE.md shaping each repository.
Choose GitHub Copilot if
- Your code and reviews live on GitHub, and you want issues assigned to an agent that returns draft pull requests.
- You want models from OpenAI, Anthropic, Google, and others on one GitHub bill, with model policies per organization.
- You need IP indemnity and Copilot policies under a GitHub agreement.
Side by side
| Dimension | Claude Code | GitHub Copilot |
|---|---|---|
| Where it runs | Terminal CLI, VS Code and JetBrains extensions, desktop app, web, and mobile12 | Extensions for VS Code, Visual Studio, JetBrains IDEs, Eclipse, and Xcode, plus GitHub.com, Copilot CLI, a desktop app, and a cloud agent89 |
| Individual priceList prices in US dollars. Included usage differs by plan and is not directly comparable. | Included in Claude Pro: US$20 per month billed monthly, or US$17 per month billed annually5 | Pro: US$10 per month; Pro+: US$39 per month6 |
| Team price | Claude Team Standard seat: US$25 per seat per month billed monthly, or US$20 billed annually5 | Copilot Business: US$19 per granted seat per month, adding 1,900 AI credits per seat to a shared pool67 |
| Enterprise price | Claude Enterprise: US$20 per seat per month billed annually, plus usage at API rates5 | Copilot Enterprise: US$39 per granted seat per month, adding 3,900 AI credits per seat to a shared pool67 |
| Models | Anthropic's Claude models15 | Models from OpenAI, Anthropic, Google, xAI, Microsoft, and others, priced per token in AI credits14 |
| Permissions | Auto mode by default, where a classifier reviews actions and blocks unsafe ones; manual mode starts read-only and asks before edits and commands3 | The CLI asks before running tools unless approval flags are set; the cloud agent runs behind a firewall and pushes only to its own branch910 |
| Project instructions | CLAUDE.md; reads AGENTS.md when no CLAUDE.md is present4 | .github/copilot-instructions.md, AGENTS.md, or a root CLAUDE.md or GEMINI.md11 |
| Pull request review | Code Review on Team and Enterprise for GitHub repositories, triggered per repository; Anthropic says reviews average US$15 to US$25 and engineers mark under 1% of findings incorrect1617 | Copilot code review, on request or automatic, and optionally for members without a Copilot license. About US$0.05 to US$1 of AI credits per review at Lite effort, plus Actions minutes. GitHub says 71% of reviews surface actionable feedback1218 |
| Agent pull requests merged | 84.3%, a statistical model's estimate from 191 pull requests in 22 Python repositories with 100+ stars, data to Nov 2025; 75.8% of all 219 merged19 | 59.6% in the same model; 69.4% of all 1,259 merged19 |
What practitioners report, side by side
Each product's figures come from its own practitioners, for the group you pick.
Overall signal
Leaders and teams rate Claude Code +72, 61 points above GitHub Copilot at +11. Most of the gap is in praise, with 42 of Claude Code's 55 sources positive against 9 of GitHub Copilot's 32. Claude Code has 3 negative and GitHub Copilot 8.
The score counts the sources from leaders and teams that take a side and is how far positive votes lead negative ones, per 100 votes. A confirmed leader casts three votes, another confirmed practitioner two and any other source one, so Claude Code's +72 rests on votes split 66 positive, 18 mixed and 3 negative and GitHub Copilot's +11 on 16, 27 and 10. A lead of a fifth of the votes reads positive and a lead of half strong positive, with the same lines below zero. The dots show each source once, positive, mixed, negative. Use a task pilot to measure accepted work and review time.
Footprint on the Agentic Software Development stack
They differ on five of 12 layers. Claude Code owns two that GitHub Copilot does not, and GitHub Copilot owns one that Claude Code does not. Coverage is each product's placement on this stack, as its product page shows it. Core is a job the product owns; Supports and Touches are jobs it helps with. Coverage says nothing about how well a product does them.
| Layer | Claude Code | GitHub Copilot |
|---|---|---|
| Build | ||
| Codebase context | Core | Supports |
| Tool execution | Core | Supports |
| Validate | ||
| Code review | Supports | Core |
| Application security | Touches | Supports |
| Ship | ||
| Source control | Touches | Supports |
The 7 layers where they match
| Layer | Claude Code | GitHub Copilot |
|---|---|---|
| Plan | ||
| Work planning | Supports | Supports |
| Build | ||
| Coding | Core | Core |
| Agent runtime | Core | Core |
| Validate | ||
| Debugging | Supports | Supports |
| Testing | Supports | Supports |
| Ship | ||
| CI/CD | Touches | Touches |
| Operate | ||
| Observability | Touches | Touches |
What practitioners praise
Among leaders and teams, Claude Code alone has four praise themes that reach 5 sources from 3 authors. Codebase Understanding and Large Refactors lead, at 9 of 9 and 8 of 9. Both carry Enterprise Administration. It is praised in 4 of the 5 Claude Code sources that discuss it, against 3 of 7 for GitHub Copilot.
| Praised for | Claude Code | GitHub Copilot |
|---|---|---|
| Enterprise Administration | 4 of 5 | 3 of 7 |
| Codebase Understanding | 9 of 9 | Not a theme in its sources |
| Large Refactors | 8 of 9 | Not a theme in its sources |
| Developer Experience | 7 of 7 | Not a theme in its sources |
| Multi-Step Execution | 5 of 6 | Not a theme in its sources |
Each count is the sources that praise a theme, or raise it as a concern, out of those that discuss it. A row pairs the two products when both carry a theme under the same or an equivalent name. A count ending in "so far" has under 5 sources from 3 authors among leaders and teams. "Not a theme in its sources" means GILD found no theme of that name in that product's sources.
What practitioners raise as concerns
Among leaders and teams, Claude Code alone has three concern themes that reach 5 sources from 3 authors. Inconsistent Results on Large Tasks and Context Management lead, at 6 of 7 and 6 of 13. Of the two themes both carry, Requires Strong Review differs most. It is raised in 15 of the 22 Claude Code sources that discuss it, against 5 of 7 for GitHub Copilot. GitHub Copilot alone has three concern themes that reach 5 sources from 3 authors. Measuring Impact and Quality Against Alternatives lead, at 5 of 13 and 3 of 7.
| Raised as a concern | Claude Code | GitHub Copilot |
|---|---|---|
| Requires Strong Review | 15 of 22 | 5 of 7 |
| Usage / Cost | 4 of 10 | 3 of 7 |
| Inconsistent Results on Large Tasks | 6 of 7 | Not a theme in its sources |
| Context Management | 6 of 13 | Not a theme in its sources |
| Permissions & Safety | 2 of 10 | Not a theme in its sources |
| Measuring Impact | Not a theme in its sources | 5 of 13 |
| Quality Against Alternatives | Not a theme in its sources | 3 of 7 |
| Data, IP & Security | Not a theme in its sources | 2 of 6 |
20 sources
- Claude Code docs: Overview
- Claude Code docs: Platforms
- Claude Code docs: Security
- Claude Code docs: Memory
- Claude pricing
- GitHub Copilot docs: Plans
- GitHub Copilot docs: Usage-based billing for organizations and enterprises
- GitHub Copilot docs: Where to use GitHub Copilot
- GitHub Copilot docs: Copilot CLI
- GitHub Copilot docs: Cloud agent risks and mitigations
- GitHub Copilot docs: Repository custom instructions
- GitHub Copilot docs: Copilot code review
- GitHub Copilot docs: Anthropic Claude agent
- GitHub Copilot docs: Models and pricing
- Claude Code docs: Model configuration
- Claude Code docs: Code Review
- Claude: Code Review for Claude Code
- GitHub Blog: 60 million Copilot code reviews and counting (Mar 5, 2026)
- arXiv: How Do AI Coding Agents Contribute to Software Development? An Empirical Study of Agentic Pull Requests (Jul 23, 2026)
- arXiv: Adoption and Impact of Command-Line AI Coding Agents, Microsoft's early 2026 rollout (Jul 1, 2026)