Guide

Claude Code vs GitHub Copilot, Cursor and other alternatives

Harrowtech.ai · last checked 13 October 2026

Six AI coding tools side by side: Claude Code, GitHub Copilot, Cursor, OpenAI Codex CLI, Gemini CLI and Aider. The table shows how each installs, where it runs, how it is paid for and whether it is open source, taken from each vendor's own pages. There is no ranking, because the hands-on tests listed further down have not been run yet.

Harrowtech.ai is independent of Anthropic, GitHub, Cursor, OpenAI and Google. Every outside fact below was read on the vendor's own pages on 13 October 2026. Where a page said nothing, this guide says nothing. Prices are as each page showed them, in dollars, and they change.

Side by side

ToolInstallWhere it runsHow it is paid forOpen source?
Claude Code (Anthropic)curl -fsSL https://claude.ai/install.sh | bash on macOS, Linux and WSL; a PowerShell or CMD script on Windows; also Homebrew, WinGet, apt, dnf or apkTerminal, VS Code (the extension also installs in Cursor), JetBrains, desktop app, web, and the Claude app for iOS and AndroidA paid Claude plan: Pro $17 a month billed yearly or $20 monthly, Max from $100 a month, Team or Enterprise. Or a Console account billed by usage. The free Claude plan does not include itNo. Its licence file reads “All rights reserved” and refers to Anthropic’s Commercial Terms of Service
GitHub CopilotCopilot CLI: npm install -g @github/copilot, Homebrew, WinGet or an install scriptGitHub, VS Code, Visual Studio, Xcode, JetBrains, Neovim, Eclipse, Raycast, the CLI, and the GitHub Copilot app, a desktop workspaceFree $0; Pro $10, Pro+ $39 and Max $100 a month; Business $19 and Enterprise $39 per seat a monthNo. The Copilot CLI comes under GitHub’s own licence, which does not give you the right to modify it
CursorDesktop app downloads for macOS, Windows and Linux (.deb, RPM, AppImage). Cursor CLI: curl https://cursor.com/install -fsS | bash on macOS, Linux and WSL, or a PowerShell command on WindowsDesktop app, terminal, and the web, where agents can be started “from your browser, phone, or tools like Slack, GitHub, and Linear”Free $0; Plus $8, Premium $30, Super $100 and Ultra $200 a month; Teams from $30 per user a month; Enterprise on custom pricing. Third-party models are billed separately at API list priceNo open-source licence is shown on Cursor’s pricing, download or docs pages
OpenAI Codex CLIcurl -fsSL https://chatgpt.com/codex/install.sh | sh on macOS and Linux, a PowerShell script on Windows, npm install -g @openai/codex or brew install --cask codexTerminal, editors (OpenAI names VS Code, Cursor and Windsurf), a desktop app and Codex WebIncluded with ChatGPT Free, Go ($8), Plus ($20), Pro (from $100), Business ($20 per user a month billed annually, $25 monthly), Edu and Enterprise. Or an API key, billed at API pricingYes, Apache-2.0
Gemini CLI (Google)npm install -g @google/gemini-cli, npx @google/gemini-cli to run it without installing, brew install gemini-cli or MacPortsTerminalFree tier with a Google sign-in (60 requests a minute, 1,000 a day) or a Gemini API key (1,000 requests a day). Usage-based billing for higher limits, or Vertex AI with a billing accountYes, Apache 2.0
Aiderpython -m pip install aider-install, then aider-installTerminal. Its site says you can also use it from within your IDE or editorAider’s pages show no price. You start it with an API key from a model provider (its examples use DeepSeek, Anthropic and OpenAI) or connect a local modelYes, Apache-2.0

Three things stand out from the table. Copilot, Cursor and Codex (through ChatGPT Free) have free plans and Gemini CLI has a free tier, while Claude Code needs a paid plan or a Console account. Codex CLI, Gemini CLI and Aider are open source; Claude Code, Copilot CLI and Cursor are not, as far as their own pages show. And for UK buyers, Anthropic says “Prices shown don’t include applicable tax” and Cursor says its prices “do not include applicable taxes or fees (e.g. VAT)”, so expect tax on top. Neither shows prices in pounds.

Claude Code vs GitHub Copilot

Anthropic calls Claude Code “an AI-powered coding assistant that helps you build features, fix bugs, and automate development tasks.” GitHub calls Copilot “Your AI accelerator for every workflow, from the editor to the enterprise.” Here is how the two compare on the points their pages cover.

Cheapest way inCopilot: Copilot Free at $0, with inline suggestions “Limited to 2000 completions per month” and an allowance of GitHub AI Credits. Copilot Student is free for verified students, and GitHub says verified teachers and maintainers of popular open source projects may be eligible for free access to Copilot Pro.
Claude Code: no free option. Pro costs $17 a month billed yearly ($200 up front) or $20 billed monthly.
What a personal plan includesCopilot: monthly GitHub AI Credits of 1,500 on Pro, 7,000 on Pro+ and 20,000 on Max (base credits plus a flex allotment). GitHub says code completions and next edit suggestions “are not billed in AI credits and remain unlimited for all paid plans.”
Claude Code: Anthropic says it “shares the same usage limits as the rest of your plan”. Max gives 5x or 20x the usage of Pro.
TeamsCopilot: Business $19 and Enterprise $39 per seat a month, with 1,900 and 3,900 AI credits per user a month. Each licence adds credits to a shared enterprise pool, and use beyond the pool is charged at $0.01 per AI credit.
Claude Code: Team standard seats are $20 a month billed yearly ($25 monthly) and premium seats $100 ($125 monthly). Enterprise is $20 per seat a month billed yearly, plus usage at API rates.
Using bothGitHub’s Copilot page says you can “Assign tasks to agents like Copilot, Claude by Anthropic, and OpenAI Codex”. Its plans page lists third-party agents as a public preview, not included on Free or Student.

Full Claude plan details are in Claude Code pricing.

Cursor

Cursor’s docs describe it as “a coding agent for building ambitious software.” Its download page says the desktop app “is available for macOS, Windows, and Linux”, and Cursor CLI “lets you interact with AI agents directly from your terminal to write, review, and modify code.”

How the plans work. Cursor’s plans are shared with Grok. The pricing page says “Every tier includes a pool of AI usage shared across Cursor and Grok”, and the allowance refreshes weekly. Plus ($8) includes Grok models only. Premium ($30) and above include Grok models but no allowance for third-party models; those stay available in Cursor at API list price, paid for by usage, and you can set spending caps. The same plans also carry features for X, such as “Post, reply, and follow”.

Models. Cursor’s docs list Claude Fable 5.1, Haiku 5.5, Opus 5.5 and Sonnet 5.5 alongside Composer 2.5 and Gemini, GPT and Grok models.

Cursor and Claude Code together. Anthropic’s overview page offers its VS Code extension with an install link for Cursor as well as VS Code, so choosing Cursor as an editor does not rule out Claude Code.

OpenAI Codex CLI

OpenAI’s README calls it a “Lightweight coding agent that runs in your terminal.” Run codex and choose “Sign in with ChatGPT” to use it with a ChatGPT plan, or use an API key, which the README says “requires additional setup.”

OpenAI’s pricing page says Codex is “included in your ChatGPT Free, Go, Plus, Pro, Business, Edu, or Enterprise plan”. For Plus and standard Business it gives estimated local messages per five-hour period rather than fixed limits, and says “Pro plans currently have no five-hour limit.” With an API key you “Pay for Codex usage based on API pricing”, and the cloud-based features such as GitHub code review and Slack are not included.

Gemini CLI

Google’s README calls it “An open-source AI agent that brings the power of Gemini directly into your terminal.” It offers three ways to sign in: a Google account (free tier of 60 requests a minute and 1,000 a day), a Gemini API key (1,000 requests a day free, with usage-based billing when you need more) or Vertex AI, aimed at enterprise teams, with “Higher rate limits with billing account.” The README gives no dollar prices.

Aider

“AI pair programming in your terminal.” Aider says it “works best with Claude 3.7 Sonnet, DeepSeek R1 & Chat V3, OpenAI o1, o3-mini & GPT-4o, but can connect to almost any LLM, including local models.” It makes a map of your codebase, commits its changes to git with commit messages, and can lint and test your code every time it makes changes. Aider’s own pages give no prices.

How to choose

These are questions to ask, not a ranking. Each can be answered from the vendor pages above.

  1. Which account do you already pay for? Copilot comes with a GitHub plan, Codex with a ChatGPT plan, Claude Code with a Claude plan, and Gemini CLI can start on a Google sign-in. Using what you have avoids a second bill.
  2. Terminal, editor or desktop app? Gemini CLI and Aider work in the terminal. Copilot lists many editors as well as a CLI and a desktop app. Claude Code, Codex and Cursor also each have a terminal tool and a desktop app.
  3. Do you need a free way in? Copilot, Cursor and ChatGPT have free plans, and Gemini CLI has a free tier. Claude Code needs a paid plan or a Console account.
  4. Does open source matter to you? Codex CLI, Gemini CLI and Aider are Apache-licensed. Claude Code and the Copilot CLI are not, and Cursor shows no open-source licence.
  5. Which models do you want? Aider connects to almost any model, including local ones. Copilot offers model choice on paid plans. Cursor lists Claude, Gemini, GPT and Grok models, with third-party models paid for at API list price.
  6. Will a team need central control? GitHub says Copilot Business offers “centralized management and Copilot policy control for organization members.” Cursor Teams and Claude Team both list central billing and administration and single sign-on. Compare the specifics on their pages.

What we will test

Vendor pages say what a tool offers, not how well it works on a real project. Harrowtech.ai plans to run these tests on each tool and add the results to this page:

  1. Setup time. From a clean Windows laptop and a clean Mac to the first working prompt, timed, following only the vendor’s own install steps.
  2. The same bug, fixed in each tool. One known bug with a failing test in a copy of a small real project. Record whether the test passes, how many prompts it took and how large the change was.
  3. Cost of a day’s work. A day of ordinary tasks on each tool’s cheapest paid plan, recording what the usage meter or bill shows at the end.
  4. Where the limits bite. How far each free or entry plan gets on the same task list before it hits a usage limit.
  5. Working in an unfamiliar codebase. Ask each tool to explain an existing project, then add a small feature that follows the project’s own conventions.
  6. Safety prompts. What each tool asks before it edits files or runs commands, and how easy that is to change.

Need coding tools set up?

Harrowtech.ai sets up coding tools from £720 (1 day), with half-day coding tools training from £360, at a fixed price agreed up front. See our prices or get in touch.

What this guide does not cover

It does not cover Windsurf, Amp, OpenCode or other tools. Until the tests above are run, it cannot say which tool writes better code. Check each vendor page before buying; plans and allowances change.

Related: What is Claude Code? · How to install Claude Code · Claude Code desktop app · Claude Code best practices

Sources

All guides: Guides.