Cursor Pricing (2026): Plans, Limits, Ultra

Cursor's pricing page looks nothing like it did a year ago. New self-serve buyers now get The X Subscription, one plan that covers Cursor, Grok and X. The usage pool refreshes every week instead of every month, and outside models such as Claude and GPT are billed on top of what you pay each month. If you open the page expecting the old "Cursor Pro for $20" setup, you'll be confused.

I use Grok, ChatGPT and Claude every day in my work, so I pay close attention to how AI tools charge for usage. This guide explains each Cursor plan in plain terms, shows where the hidden costs are, and compares Cursor with GitHub Copilot, plain VS Code and Claude Code, so you can choose a plan without paying for more than you need.

Every price below was rechecked on November 7, 2026 on the official pages: Cursor pricing, The X Subscription pricing docs, GitHub Copilot plans and Claude pricing. Prices are in US dollars and don't include taxes such as GST or VAT. Plans change often, so confirm before you subscribe.

Quick answer: Cursor (via The X Subscription) costs $0 (Free), $8 (Plus), $30 (Premium), $100 (Super) or $200 (Ultra) a month. Annual billing is cheaper: Plus $80/yr, Premium $300/yr, Super $1000/yr, Ultra $2000/yr. Most working developers should start on Premium. Use Plus only if Grok models are enough and you want X Premium perks. Move to Super or Ultra after you've hit Premium's weekly pool more than once. SuperGrok is not sold to new customers — new buyers pick The X Subscription plans above. Claude and GPT inside Cursor are billed separately at API list price.

Cursor pricing at a glance

Here's the full individual lineup from Cursor's pricing page and docs.x.ai:

Cursor labels Premium "Most popular." Its own advice is Premium for new AI users, Super for daily agent users, Ultra for power users, and Plus for social creators with limited AI usage.

What is The X Subscription?

The X Subscription is the current self-serve product that brings Grok, Cursor and X under one bill. According to docs.x.ai:

• New self-serve customers buy The X Subscription plans (Free through Ultra).
• Existing Cursor, SuperGrok and X Premium plans continue. Switching is optional.
• SuperGrok is not sold to new customers. Don't treat it as a plan you can still buy on the public pricing page.
• Paid individual plans include X Premium benefits when you connect your X account.
• Subscriptions are sold only through X Subscription Corp / the official site. Cursor's FAQ says purchases from resellers or third parties are unauthorized and may be suspended.

How this differs from old Cursor Pro

Older marketing talked about a standalone Cursor Pro around $20. That framing is gone for new buyers. Today you pick a tier on The X Subscription, get a shared weekly AI pool across Cursor and Grok, and treat Claude/GPT usage in Cursor as a separate usage bill. You also get X and Grok features on the same seat (with Grok Bot starting at Premium, not Plus).

If you still hold an older Cursor or SuperGrok plan, it keeps working until you change or cancel it. When you switch, eligible unused time is typically refunded prorated to your original payment method, and the new plan starts immediately — confirm the current switch rules on docs.x.ai before you move.

How Cursor's usage limits work

This is the part most people skip, and it decides how far your money goes.

One shared pool, refreshed weekly

Every tier includes a pool of AI usage shared across Cursor and Grok. Within Grok, that pool covers Chat, Imagine, Voice and Bot. Higher tiers include larger pools, and usage cost varies by product. The allowance refreshes weekly, so one heavy day can't wipe out a whole month.

When you run out, you can buy top-offs (Additional Usage credits) or wait for the next refresh. Your remaining usage is shown in settings.

Third-party models cost extra

This is the biggest thing to understand about Cursor pricing in 2026:

• Plus includes Grok models only.
• Premium and above include generous Grok usage but no included allowance for third-party models.
• Claude, GPT and other non-Grok models stay available in Cursor at API list price. You pay for what you use, and you can set spending caps.

In practice, your real monthly cost depends on which models you pick:

• Grok models only: the price on the plan is the price you pay (plus tax).
• Mostly Claude or GPT inside Cursor: the plan fee, plus a usage bill that grows with long agent sessions and big files.

The Grok API is separate too

Paid plans include Grok API access, but API calls are billed at standard model rates and don't come out of your plan's pool. That's useful if you build your own tools (new to the idea? see what an API is), but it's an extra line on the bill, not a free perk.

Grok Bot is not on Plus

Grok Bot is included on Premium, Super and Ultra. Plus does not include it. Bot usage draws from that seat's shared weekly pool, not a separate additive pool.

Which Cursor plan suits you?

Here's how I'd match plans to typical users.

Free: trying Cursor out

Use Free to see whether an AI-first editor fits the way you work. The request limits are low, so it isn't meant for real project work. You still get basic X posting, replies and follows.

Plus ($8): light coding with Grok, or social creators

Plus suits students, hobbyists and anyone who codes a few hours a week and is happy using Grok models. Docs.x.ai and Cursor both frame Plus as a fit for social creators with limited AI usage, since you also get X Premium-style perks when you connect X. The catch is no third-party models, no cloud agents, no Bugbot and no Grok Bot.

Annual Plus at $80/year (~$7/month) is a small saving if you'll keep the seat for most of a year.

Premium ($30): the default for working developers

If you write code for clients or for your own product every week, start here. You get 4x the Plus usage, cloud agents that work in the background, Bugbot reviews, Origin code hosting, Grok Bot, video generation in Grok, and the option to call Claude or GPT when a task needs them (billed separately). For most freelancers and small agencies, this is the plan that pays for itself.

Annual Premium is $300/year (~$25/month).

Super ($100): daily agent users

Super makes sense when you run agents for hours every day and keep hitting Premium's weekly pool. You get 15x Plus usage and priority model access at peak times. Annual Super is $1000/year (~$84/month).

Ultra ($200): power users

Ultra gives you 50x Plus usage and earliest access to new products and models. It's for people who run several agents in parallel all day. Most individual developers won't need it. Annual Ultra is $2000/year (~$167/month).

Teams and Enterprise

Team seats are available on Premium ($30), Super ($100) and Ultra ($200) per seat. Each seat matches the individual tier's usage. You can mix seat types on one team (for example three Premium seats and one Super), so a designer who barely codes doesn't need the same seat as your lead developer. Plus is not available for teams.

Teams adds central billing, SSO, usage analytics, team-wide privacy settings and a marketplace for internal rules, skills and plugins. Enterprise adds invoice/PO billing, SCIM, audit logs, granular spend controls and stronger admin controls for Grok Bot and network access.

Cursor vs Copilot

GitHub Copilot is the comparison most people make. Here are its current individual plans from GitHub's plans page:

Copilot runs on GitHub AI Credits, where 1 credit = $0.01. Chat, agent mode, code review, the cloud agent and Copilot CLI use credits. Code completions and next edit suggestions don't use credits, and they're unlimited on every paid plan. GitHub notes that the "flex" part of the allowance can change over time. Business and Enterprise plans exist for orgs; this guide stays focused on individual pricing.

Side-by-side

Where Copilot wins

• You keep your editor. Copilot works in VS Code, Visual Studio, JetBrains IDEs, Xcode, Neovim, Eclipse, Zed and more. Cursor is its own editor.
• Model choice is built into the price. Copilot's credits cover models from Anthropic, OpenAI, Google, xAI and others. In Cursor, non-Grok models cost extra.
• It lives inside GitHub. You can assign an issue to Copilot and get a pull request back, plus PR reviews on github.com.
• Lower entry price for paid completions. At $10, Copilot Pro gives you unlimited completions.

Where Cursor wins

• The editor is built around AI. Agent work, diffs and review all happen in one place instead of in a plugin.
• More bundled usage for Grok users. If Grok models handle most of your work, Cursor's weekly pool goes a long way.
• Extras in one subscription. Premium includes cloud agents, Bugbot reviews and code hosting, along with Grok's chat and media tools — and X Premium benefits when you connect X.

Verdict on Cursor vs Copilot: if you want Claude and GPT models without a separate usage bill, or your team can't switch editors, Copilot is the simpler buy. If you're happy to work mainly with Grok models and want an editor designed around agents, Cursor Premium is the stronger daily tool. For more options, see Copilot alternatives and Cursor alternatives.

Cursor vs VS Code

People searching cursor vs vscode usually want to know whether it's worth leaving a free editor.

Cursor is built on the VS Code codebase, so the layout, shortcuts and most extensions will feel familiar, and moving over is usually quick. The real difference is cost and how AI fits in:

• VS Code is free and open source. On its own it has no paid AI. You add AI through extensions, most often GitHub Copilot, which costs $0 to $100 a month depending on the plan.
• Cursor puts AI at the center of the editor. You pay for a plan, and the agent, chat and review features are part of the editor rather than add-ons.

A sensible way to decide:

• Stay on VS Code if you mostly need autocomplete and occasional chat, if your company standardizes on VS Code, or if you rely on an extension that doesn't behave the same way in Cursor.
• Move to Cursor if you hand larger tasks to an agent most days and want that workflow to feel native.

You can also try both. Install Cursor next to VS Code, open the same project and use each for a week on real work before you pay for anything. Closest agent-first rival: Windsurf vs Cursor.

Claude Code vs GitHub Copilot: a short note

Claude Code is a different kind of tool. It's an agent you run mainly from the terminal, and you give it a whole task instead of working line by line in an editor. According to Claude's pricing page, it's included with every paid Claude plan:

• Free: $0 — Claude Code not included
• Pro: $20/month, or $17/month billed annually ($200 up front)
• Max: $100/month — 5x Pro usage every 5 hours and every week; $100/month API credits
• Ultra: $200/month — 10x Pro weekly usage, no 5-hour limit for verified accounts, $200/month API credits

Claude Code shares one usage pool with Claude chat. Limits reset in rolling five-hour windows (except Ultra for verified accounts), and paid plans also have weekly limits. Our Claude Code pricing guide explains those limits in detail; full Claude chat pricing lives in Claude pricing.

On Claude Code vs GitHub Copilot, Copilot is the better everyday assistant inside your editor and on GitHub, and it starts at $10. Claude Code is better when you want to hand off a multi-file job, like adding tests to a module or refactoring a feature, and then review the result. Copilot can now pass some tasks to third-party agents, including Claude, in preview on higher plans, so the two can work together instead of competing.

For a deeper side-by-side of the three agent tools, read Cursor vs Claude Code vs Codex, or the head-to-head Claude vs Copilot if those are your two finalists.

Is Cursor worth it in 2026?

For most working developers, yes, as long as you pick the plan based on how you'll actually use it.

• Founders building an MVP: Premium at $30, or Claude Pro at $20 if you prefer to delegate whole tasks to Claude Code.
• Freelancers and agency developers: Premium is the right starting point. Set a spend cap before you use Claude or GPT models.
• Students and hobbyists: Plus at $8 if Grok models work for you, or Copilot Free if you'd rather stay in VS Code.
• Social creators who code lightly: Plus, per Cursor's own recommendation.
• Teams that live on GitHub: Copilot Pro or Pro+ is usually less disruptive, with Cursor Teams as the choice if your developers want an agent-first editor.
• Heavy agent users: move up to Super only after you've hit Premium's weekly limit more than once; Ultra only if Super still isn't enough.
• Grok + X users who also code: The X Subscription is the cleanest way to get both on one bill — see also Grok vs ChatGPT if you're weighing chat products.

The mistake to avoid is buying a large plan up front. Start one tier lower than you think you need, watch the usage meter for two weeks and upgrade only when the limit actually slows you down. If none of these plans fit your budget, my Cursor alternatives roundup covers cheaper and free options, Windsurf vs Cursor compares the closest rival plan by plan, and Replit alternatives covers browser AI builders if you don't want a local editor at all.

FAQ

Final thoughts

Cursor pricing in 2026 is simple at the top — five personal plans from $0 to $200 under The X Subscription — but your real cost depends on the models you use and how hard you push the weekly pool. Premium is the plan I'd point most developers to first. Copilot is the easy choice if you want to keep your current editor and multi-provider models in one credit pot. Claude Code is worth adding when you want to delegate bigger terminal tasks (and Ultra on Claude removes the five-hour cap for verified accounts).

Pricing in this space changes often, so check cursor.com/pricing and docs.x.ai/x-subscription/pricing before you subscribe. If you'd like help choosing an AI coding setup for your project, you can learn more about me and my work here.