Disclosure: some links on this page may be affiliate links; we may earn a commission at no extra cost to you. Read the disclosure.

By the toolist.ai editors. Researched from Cursor's own pages and hands-on tests published by others, not hands-on tested by toolist.ai. Updated October 16, 2026.

Cursor review: pricing, features and what testers found

Cursor is an AI code editor built on VS Code, with an agent that edits across your codebase, Tab autocomplete and cloud agents that open pull requests. Since October 12, 2026 new customers buy it as part of The X Subscription, with Grok and X Premium: Plus at $8 a month runs only Grok models, and Premium at $30 is the cheapest plan with cloud agents, Bugbot and third-party models.

Our take

Cursor suits developers who want an AI agent inside a familiar VS Code-style editor: Zapier found its agents still ahead of GitHub Copilot's, and TechSifted estimated routine coding went 35 to 40% faster over a month, though about 15 to 20% of agent outputs needed fixing. Since October 12, 2026 new customers buy it as The X Subscription, bundled with Grok and X Premium. Plus at $8 a month runs only Grok models, and Cursor itself recommends it for social creators and Premium at $30 for new AI users; Premium is the cheapest plan with cloud agents, Bugbot and Claude, GPT or Gemini, which are no longer included and cost API list price from prepaid credit. Usage is a weekly pool shared with Grok that Cursor states only in relative terms, such as multiples of Plus's. Existing subscribers can keep their legacy Pro, Pro+, Ultra or Teams plan, which still includes third-party model usage.

Based on Cursor's published plans and the hands-on tests linked below, not on a test of our own.

Good for

  • Professional developers doing multi-file changes and refactors in an editor they know from VS Code
  • Handing tasks to background cloud agents that come back with a pull request, on Premium and above
  • People who already pay for Grok or X Premium and want Cursor in the same subscription

Watch out for

  • Plus at $8 runs only Grok models, at fixed settings (Fast off, medium effort), with no Auto, cloud agents, Automations, Bugbot, third-party models or your own API keys.
  • Third-party models such as Claude, GPT and Gemini are not included on any new plan: from Premium up they cost API list price from prepaid credit (top-ups), Teams pay $0.25 per million tokens on top, and Plus and Free cannot use them.
  • Cursor states usage only in relative terms, such as multiples of Plus (Premium 4 times, Super 15, Ultra 50), never as an amount of work; the pool refreshes weekly, is shared with the Grok apps and Grok Bot, and unused usage does not roll over. Past it you wait for the reset or pay the same per-model rates from prepaid credit, which expires 365 days after purchase.
  • New plans are sold by X Subscription Corp and billed at grok.com under SpaceXSI's subscription terms and privacy policy: payments are final unless the law requires a refund, you must cancel at least 24 hours before renewal, and the terms let SpaceXSI bring your Cursor account data, including coding data, into your SpaceXSI account when you sign in with Cursor.
  • Existing subscribers keep their legacy plan, price and features until they cancel or switch: Hobby, Pro ($20 a month), Pro+ ($60), Ultra ($200) and Teams Standard ($40 a user) or Premium ($120) seats, with monthly pools that include third-party models (except on Hobby) and, on Pro, Pro+ and Ultra, Cursor's 14-day refund rule. Switching is permanent and drops the included third-party usage; Cursor maps Pro to Premium and Pro+ to Super.
  • With Privacy Mode off, Cursor may store your code and prompts and use them to train its models (team members have it on by default). Cursor says Privacy Mode applies in the Cursor editor and that Grok Bot has its own data flows, and SpaceXSI says it may use your content and interactions with Grok to train its models, which you control with its Improve the model setting.

What Cursor does

Price

Free plan: Free gives limited use of Agent, Chat and Tab completions with Cursor's own Grok and Composer models, without third-party models, cloud agents or Bugbot; Cursor does not publish the exact limits. Existing users of the legacy Hobby plan keep it, and it works the same way.

PlanBilled monthlyBilled yearly, per monthIncludes
Plus $8 $6.67 Base weekly usage pool shared with Grok; Cursor's Agent with Grok models only (Grok 4.7, 4.6 and 4.5, plus Composer 2.5) at fixed settings, without Auto, cloud agents, Automations, Bugbot, Origin, third-party models or your own API keys; Grok Chat, Voice, Imagine, Build and API access; X checkmark and fewer ads
Premium $30 $25 Everything in Plus with 4 times Plus's usage, the latest Grok models, cloud agents and Automations, Origin code hosting, Bugbot code reviews, Grok Bot and third-party models at API list price from prepaid credit; ad-free X on most surfaces
Super $100 $83.33 Everything in Premium with 15 times Plus's usage and priority model access at peak times
Ultra $200 $166.67 Everything in Super with 50 times Plus's usage and earliest access to new products and models
Teams $30 not listed Per user: Premium ($30), Super ($100) or Ultra ($200) seats with the matching plan's weekly usage, mixed on one team (no Plus seats); team billing and admin, team marketplace for rules, skills and plugins, usage analytics, team-wide Privacy Mode, custom Chat data retention, SAML/OIDC SSO; third-party models at API list price plus $0.25 per million tokens from a shared prepaid balance; no yearly price published
Enterprise not listed not listed Contact sales: usage pooled across Grok and Cursor, invoice/PO billing, SCIM, granular spend controls, repository, model and MCP access controls, audit logs and service accounts, priority support

Prices last checked: 2026-10-16 on Cursor's pricing page. Full breakdown: Cursor pricing.

Price history

  • October 16, 2026: New plan: Plus ($8 a month billed monthly, $6.67 billed yearly). New plan: Premium ($30 a month billed monthly, $25 billed yearly). New plan: Super ($100 a month billed monthly, $83.33 billed yearly). Ultra, billed yearly: $160 → $166.67 a month. New plan: Teams ($30 a month billed monthly). Plan dropped: Pro. Plan dropped: Pro+. Plan dropped: Teams Standard. Plan dropped: Teams Premium.

First recorded October 9, 2026. We re-check prices every week and list each change here.

What hands-on testers found

We have not tested Cursor ourselves. These are the published tests we found that used it hands-on, newest first; follow the links for the full reviews.

Zapier · Miguel Rebelo · updated June 8, 2026

What they tested: Cursor side by side with GitHub Copilot on codebase context, multi-file editing, agents, GitHub integration and price, as part of a round of testing AI coding tools.

  • Cursor's Tab autocomplete draws on an index of the whole project, which Copilot's does not, so suggestions fit the codebase more often.
  • Its agent tools, such as parallel agents and Cursor 3's agents window, still lead Copilot, though by a smaller margin than before.
  • Linking agents to GitHub issues and CI takes extra setup, and at $20 for Pro, before the October 2026 change, it cost twice Copilot Pro, so it pays off mainly for regular multi-file agent work.

TechSifted · Marcus Rowe · April 3, 2026

What they tested: Thirty days as his only editor on the $20 Pro plan, before the October 2026 change made it a legacy plan, building a Next.js SaaS dashboard, a Python data pipeline and a Chrome extension.

  • He accepted about 70 to 78% of Tab suggestions and estimated routine work went 35 to 40% faster.
  • Agent mode moved a 14-file auth system to NextAuth.js with three manual fixes, but about 15 to 20% of agent outputs needed corrections.
  • He ran out of Pro usage around day 22 and bought more twice (on Pro's monthly allowance, before the October 2026 change), and a 200,000-line monorepo made the editor lag.

Their score: 4.3 out of 5

Unite.AI · Janine Heinrichs · updated March 30, 2026

What they tested: A habit-tracking web app built from one prompt in a new project, then a follow-up request to restyle a button.

  • Cursor wrote the app's files from one prompt, and the app ran in the browser without manual edits.
  • The button color change was applied within seconds, with an explanation of what it did.
  • She judged it better suited to experienced developers on complex projects, and warns it can produce wrong code and slow a computer down.

What users say

Test result

toolist.ai has not run its own test of Cursor; the findings above come from the published tests linked. See how we research and test.

Best coding tools · Cursor pricing

Do you make Cursor? Claim this listing to correct it; it's free. To see whether ChatGPT, Claude, Gemini and Perplexity recommend it, and how many readers we send you: toolist.ai for AI companies.

Sources