Claude Code vs Cursor: same $20, different shape

A disclosure before anything else: this text is drafted by a model built by Anthropic, which makes one of the two products compared here. That is a conflict, and the way we handle it is to compare only what both vendors publish, to rank nothing on quality, and to link every claim to the page it came from. If you want a verdict on which writes better code, this is not that article and it could not honestly be written here.

What each one is

Cursor is an editor. You install it, open your project in it, and the agent works inside that window. The plan page lists Composer on the free tier, and MCPs, skills, hooks and cloud agents from Pro upward.

Claude Code is an agent that runs where you already work. Its documentation lists the terminal, the IDE, the desktop app and the browser, with extensions for Visual Studio Code and JetBrains, plus code review and CI through GitHub Actions or GitLab CI/CD. It also speaks MCP, packages repeatable work as skills like /review-pr or /deploy-staging, and runs shell commands before or after its own actions through hooks.

Line the feature lists up and they mostly rhyme. The real question is whether you want to change editors.

What each one costs

Cursor: Hobby is free with limited agent requests. Individual Pro is $20 a month. Pro+ carries three times Pro's agent limits and Ultra carries twenty times; the pricing page names those multiples without publishing their prices. Teams is $40 per user a month, with Premium at five times Standard's limits. Enterprise is custom.

Claude Code comes inside Claude Pro, which is $20 billed monthly or $17 a month on an annual subscription billed at $200 up front. Max starts at $100 a month and offers five or twenty times Pro's usage, higher output limits, early access and priority at busy times.

Both meter usage rather than seats alone. Cursor states that every plan includes a set amount of model usage and that on-demand usage past it is billed in arrears. Claude's plans carry usage limits too, and Claude Code can also be run against the API on pay-per-token pricing, which is the route worth costing out if your usage is bursty rather than daily.

• Free entry: Cursor Hobby, with limited agent requests.
• $20 a month: Cursor Pro, or Claude Pro billed monthly.
• $17 a month: Claude Pro on an annual subscription, $200 up front.
• Scaling up: Cursor Pro+ at 3x and Ultra at 20x; Claude Max from $100 at 5x or 20x.
• Teams: Cursor at $40 per user a month; Claude team plans go through sales.

What decides it in practice

If your team already lives in one editor and will not move, the agent that runs in the terminal and attaches to that editor costs you nothing in habit. If your team is happy to adopt a new editor and wants the agent woven through it, that is the case for Cursor and it is a real one.

If the work is mostly automation — review on every pull request, jobs in CI, scripted passes over a repository — check what each side does without a human at the keyboard. Cursor lists cloud agents and Bugbot for agentic code review, billed by usage. Claude Code documents GitHub Actions and GitLab CI/CD paths.

If procurement matters, read the top of both lists. Cursor's Enterprise tier publishes SCIM seat management, audit logs, service accounts, and repository, model and MCP access controls. Claude publishes SSO, SCIM, audit logs, role-based access and a HIPAA-ready offering on its higher plans.

What this comparison refuses to do

It does not tell you which produces better code. Doing that honestly means running both on your repository, with your tests, over enough tasks that the result is not noise — and then the answer is about your codebase, not about the products.

It also does not price Pro+ and Ultra, because Cursor's pricing page names their limits without naming their prices at the time of writing. If those tiers are your target, get the number from Cursor before you plan around it.