GitHub Copilot vs Claude Code.
GitHub’s terminal agent against Anthropic’s. Both can run Claude models: Copilot bills them in AI credits at per-token rates, and a Claude plan meters them against usage windows. Every figure below was read on 16 November 2026 from the page linked under Sources. Starkomand, which runs both, is ours; the comparison does not depend on it.
v1.0.8 · Mac and Windows · free for 3 days · no account
The short version.
- PriceCopilot Pro is $10 a month and includes $15 of AI credits. Claude Pro is $20 a month and includes Claude Code.
- Claude in CopilotCopilot CLI runs Claude models at the same per-token rates as Anthropic’s API, drawn from your credits: Sonnet 5.5 at $2 and $10 per million tokens in and out, Opus 5.5 at $4 and $20. Opus needs Pro+ or higher.
- Other modelsCopilot also offers GPT, Gemini and Grok models, or your own provider. Claude Code runs Claude models.
- DefaultsCopilot CLI asks before each tool, with an opt-in local sandbox and cloud sandboxes. Claude Code starts in auto mode, where a classifier reviews actions, with an opt-in sandbox.
- InstructionsCopilot CLI reads
CLAUDE.mdandAGENTS.mdas well as its own files, so one set of rules can serve both.
Side by side.
GitHub Copilot CLI next to Claude Code, from both vendors’ public pages on 16 November 2026.
| Criterion | GitHub Copilot CLIGitHub | Claude CodeAnthropic |
|---|---|---|
| Where it runs | The terminal, with Copilot also in VS Code, JetBrains, Xcode, Zed and on GitHub | The terminal, VS Code, JetBrains, a desktop app and the web |
| Models | Claude, GPT, Gemini and Grok models, or your own provider | Fable 5.2, Opus 5.5, Sonnet 5.5 and Haiku 5.5Fable is included on Max and paid with usage credits on Pro. |
| How usage is billed | GitHub AI Credits at per-token rates for each modelOne credit is $0.01. | Usage windows on a plan, or per token on an API key |
| Cheapest plan with the agent | Copilot Free, with limited chat and agent usage; Pro at $10/month | Claude Pro, $20/monthClaude’s Free plan does not include Claude Code. |
| Top individual plan | Copilot Max, $100/month with $200 of credits | Claude Ultra, $200/monthNo five-hour limit for verified accounts. |
| Starting permissions | Asks before each tool; /permissions assisted lets a model approve safe requests |
Auto mode: a second model, the classifier, reviews actions instead of youSince v2.1.283. Manual mode asks before most edits and commands. |
| Sandbox | Opt-in: /sandbox enable locally, or --cloud for a GitHub-hosted sandbox |
Opt-in with /sandboxOn macOS, Linux and WSL2. |
| Context window | 1M tokens on the latest models, chosen per sessionThe larger window uses more credits. | 1M tokens on every current model, at standard API pricing |
| Parallel work | /fleet subagents, and several named sessions |
Subagents, and experimental agent teams |
| Project instructions | .github/copilot-instructions.md, AGENTS.md, CLAUDE.md and GEMINI.md |
CLAUDE.md |
| Terminal-Bench 2.1, best published run | No published run | 83.8% with Claude Fable 5 |
Same on both. MCP servers, a plan mode before edits, and your own Anthropic API key as an option.
Every figure on this page was read on 16 November 2026 from the page listed under Sources, and both products change monthly; product names belong to their owners. Spotted a change? Tell us.
Prices.
Copilot sells credits for every model; Anthropic sells usage windows for Claude.
| Plan | Price | What it includes |
|---|---|---|
| Copilot Free | $0 | Limited chat and agent usage, Copilot CLI included |
| Copilot Pro | $10/month | $15 of AI credits a month: $10 base and $5 flex |
| Copilot Pro+ | $39/month | $70 of AI credits a month; premium models including Opus |
| Copilot Max | $100/month | $200 of AI credits a month; priority access to new models |
| Claude Pro | $20/month, or $17/month billed annually | Claude Code; Fable paid with usage credits |
| Claude Max | $100/month | Five times Pro’s usage; Fable included, up to half the weekly limit |
| Claude Ultra | $200/month | No five-hour limit for verified accounts; ten times Pro’s weekly usage |
Copilot’s flex allotment is variable additional usage that GitHub says may change over time. Past the included credits, a budget you set keeps Copilot going at $0.01 a credit.
At the API
Claude models in Copilot against Anthropic’s API, per million tokens of input and output.
| Model | In Copilot | Anthropic API |
|---|---|---|
| Claude Fable 5.2 | $8 / $40 | $8 / $40 |
| Claude Opus 5.5 | $4 / $20 | $4 / $20 |
| Claude Sonnet 5.5 | $2 / $10 | $2 / $10 |
| Claude Haiku 5.5 | $0.10 / $0.50 | From $0.10 / $0.50 |
Copilot charges Anthropic’s API rate, so the plan’s credits are a prepaid amount at list price. In our own logs the median Claude Code session cost $2.38 at list price (see what a session costs), so Pro’s $15 of credits covers about six sessions like it.
A Claude plan meters the same work against five-hour and weekly windows instead. For daily Claude work that is the cheaper route; Copilot’s credits suit occasional Claude use next to GitHub’s other models.
Which one.
Pick Copilot if
you live on GitHub, want one bill for several model families, or want a $10 way in with credits you can stretch on lighter models.
Pick Claude Code if
you use Claude models all day, since a flat plan meters them by window rather than per token, or you want auto mode from the first run.
Pick both if
you have Copilot through work and Claude for yourself, or want Copilot’s GPT and Gemini models next to Claude Code. Both read the same CLAUDE.md.
MoreGitHub Copilot CLI in StarkomandClaude Code in StarkomandCodex vs Claude CodeCursor vs Claude CodeOpenCode vs Claude CodeCodex vs CursorGemini CLI vs Claude CodeRun multiple Claude Code agents in parallelWhat is an ADE?Agent orchestrationcmux for Windows
Running both by hand.
Copilot CLI reads CLAUDE.md, so the rules you already wrote for Claude Code apply to both with no extra file.
# a worktree each when both write code
git worktree add ../app-copilot -b copilot-task
# then, in a terminal each
claude
cd ../app-copilot && copilot
Give the two agents different jobs, and when both write code, give each its own git worktree so neither overwrites the other’s changes.
Running both in Starkomand.
Starkomand is ours, so this is the one section about it. It runs on Apple silicon Macs and on Windows 10 and 11, and needs no Starkomand account: Copilot CLI and Claude Code each run in a real terminal window on one canvas, signed in with your own plans or keys.
- Plan in one, build in the other
- Open Agent actions in a window, choose Hand off… and pick the other agent. The brief carries the objective and acceptance criteria, and the new window lands wired to the one that wrote it.
- See both allowances
- Settings → Usage meters puts a ring in the rail for a metered Copilot plan and one for Claude, each with its reset.
- Read across windows
- An agent can read another window’s transcript by name, so the plan does not need retyping.
- Redirect by voice
- Hold Option and speak to an agent by name.
We have not measured whether this makes anyone faster. What it changes is that each agent stays on its own plan, in view, and work moves between them without retyping.
Questions
Is GitHub Copilot better than Claude Code?
There is no shared benchmark: Terminal-Bench 2.1 lists no Copilot CLI run, and Claude Code’s best is 83.8% with Claude Fable 5. Copilot CLI can run the same Claude models, so for a Claude model the difference is the harness and the bill.
Can Copilot use Claude models?
Yes. Copilot CLI offers Claude models including Sonnet 5.5 and Opus 5.5, billed in AI credits at the same per-token rates as Anthropic’s API. Opus needs Pro+ or higher.
Which is cheaper, Copilot or Claude Code?
Copilot Pro costs $10 a month against $20 for Claude Pro, but its $15 of credits is spent per token. In our logs a median Claude Code session cost $2.38 at list price, so $15 covers about six; a Claude plan meters the same work against usage windows instead.
Does Copilot CLI work on the free plan?
Yes, with limited chat and agent usage.
Can I use my own Anthropic API key in Copilot CLI?
Yes. Set COPILOT_PROVIDER_TYPE to anthropic and name the model with COPILOT_MODEL; copilot help providers lists the details.
Does Copilot CLI read CLAUDE.md?
Yes. It reads CLAUDE.md, AGENTS.md and GEMINI.md in the standard locations, as well as .github/copilot-instructions.md.
Try it against your own setup.
Every feature, every provider, free for 3 days. Your CLIs are already installed; the canvas takes a minute.
v1.0.8 · Mac and Windows · no account