- Cursor is now sold as part of The X Subscription: Free $0, Plus $8, Premium $30, Super $100, Ultra $200 a month.
- Paid plans include a weekly pool for Cursor's own models, shared with Grok.
- Claude, GPT and Gemini are not included: from Premium up you pay API list price from prepaid credit.
- There is no standing student discount.
Cursor now costs between $0 and $200 a month: Free $0, Plus $8, Premium $30, Super $100 and Ultra $200, before tax. Paid plans include a weekly pool for Cursor’s own models, shared with Grok. Claude, GPT and Gemini sit outside the pool, and there is no standing student discount.
Cursor is a code editor with a built-in AI agent that reads your project, edits files and runs commands. If you are new to this, start with what an AI agent is and what vibe coding means. In this comparison I pull the prices from Cursor’s pricing page and docs as they read today, explain what you pay for on each plan, and work out what one sample session costs on different models. At the end I say which plan I would start with.
What changed in Cursor pricing
Cursor is no longer sold on its own. According to the Models & Pricing docs, it is part of The X Subscription, a single subscription for Grok, Cursor and X. That follows the acquisition: on August 14, 2026, Cursor announced on its blog that it is now part of SpaceX.
The biggest change is what the price covers. A paid plan gives you one weekly pool for Cursor’s own models: Grok 4.7, Grok 4.6, Grok 4.5 and Composer 2.5. The Grok apps draw from that same pool too: Chat, Imagine, Voice and Grok Bot. Third-party models such as Claude, GPT and Gemini are not included in any of the new plans.
Older plans, including Pro at $20, keep working for people who already have them. I cover them in their own section below.
How much does Cursor cost? Pricing in one table
The cursor.com/pricing page lists five individual plans. They differ in the size of the pool and in whether you can use third-party models.
| Free | Plus | Premium | Super | Ultra | |
|---|---|---|---|---|---|
| Monthly price | $0 | $8 | $30 | $100 | $200 |
| Yearly billing | n/a | $80 | $300 | $1,000 | $2,000 |
| Weekly pool | limited usage | base (1×) | 4× Plus | 15× Plus | 50× Plus |
| Agent in the editor | yes, limited | yes, Cursor models only | yes | yes | yes |
| Claude, GPT, Gemini | no | no | at API list price | at API list price | at API list price |
| Cloud agents and automations | no | no | yes | yes | yes |
| Bugbot and Origin | no | no | yes | yes | yes |
| Extras | X account | Grok, Grok API access | Grok Bot | priority access at peak times | earliest access to new releases |
Last checked: November 21, 2026, prices from cursor.com/pricing and the docs as they read that day.
Paying yearly saves roughly two months: Premium is $300 instead of 12 × $30, which is $360. Prices exclude tax and are shown in US dollars. There is one more note under the plans: you can only buy the subscription directly, through X Subscription Corp. Cursor treats accounts bought from resellers as unauthorized and may suspend them.
Can I use Cursor AI for free?
Yes, with tight limits. The Free plan gives you Agent and Chat with limited usage, but no third-party models, no cloud agents and no Bugbot code review. Cursor does not say exactly how many requests the limit allows.
Plus at $8 is the entry plan. Its agent runs only on Cursor’s own models, at fixed settings: Fast mode off, medium effort, and Grok 4.7 with a 256k-token context window. The docs also list what Plus leaves out: Auto mode, cloud agents, automations, Bugbot, Origin, third-party models and your own API keys. Cursor itself recommends Plus for social creators with light AI use, and Premium for people new to the agent.
Does Cursor have a student discount?
Not a standing one. On its students page, Cursor says anyone can start for free, and that better access comes through promotions at campus and online events starting this fall. A campus newsletter, which you can join on the same page, announces those events and discounts.
So today a student has three routes: the Free plan, Plus at $8, or a promotion from an event if one comes along. I would steer clear of discount codes from resellers: as noted above, Cursor only sells subscriptions directly.
How the weekly pool works, and what happens when it runs out
The pool refreshes every week, and unused usage does not roll over. According to the usage and limits help page, you can see how much you have used and the next reset date in your editor settings and under Usage at grok.com/dashboard. Different models draw from the pool at different rates, and Cursor does not express the pool in tokens or dollars, only as a multiple of Plus.
When the pool runs out you have three options: wait for the reset, upgrade, or keep working from prepaid credit. You top up by hand or turn on Auto Top-up from a threshold you choose. Separately, you set a monthly additional usage limit that caps spending regardless of your balance. Each top-up stays valid for 365 days after purchase. The docs also state that requests are never downgraded in quality or speed when the pool is empty.
Prices for Cursor models beyond the pool, per million tokens:
| Grok 4.7 | Grok 4.7 (Fast) | Composer 2.5 | |
|---|---|---|---|
| Input | $2 | $4 | $0.50 |
| Cache read | $0.50 | $1 | $0.20 |
| Output | $6 | $12 | $2.50 |
Claude, GPT and Gemini in Cursor: what they cost
Third-party models are not in the pool of any new plan. From Premium up you can use them from the very first request at API list price, paid from prepaid credit. Plus and Free do not offer them at all. On individual plans Cursor adds no fee of its own on top. It does on Teams and Enterprise: $0.25 per million tokens on eligible models.
| Claude Opus 5.5 | Claude Sonnet 5.5 | GPT-5.6 Terra | Gemini 4 Neon | |
|---|---|---|---|---|
| Input | $4 | $2 | $2 | $3 |
| Cache read | $0.20 | $0.10 | $0.20 | $0.15 |
| Output | $20 | $10 | $12 | $15 |
Last checked: November 21, 2026, prices from the table in the Models & Pricing docs.
Plus does not allow your own API keys. According to the usage and limits help page, on individual plans requests made with your own key are billed directly by the model maker and use neither your pool nor your Cursor credit.
A worked example: one session on different models
To see the difference, I worked out the cost of one sample agent session. The numbers are my assumption, not a measurement: 100k tokens of new input, 900k tokens read from the prompt cache (the agent resends the whole conversation at every step) and 50k output tokens. I use the prices from the tables above and leave out cache writes.
| Composer 2.5 | Claude Sonnet 5.5 | Grok 4.7 | GPT-5.6 Terra | Claude Opus 5.5 | |
|---|---|---|---|---|---|
| Session cost | $0.36 | $0.79 | $0.95 | $0.98 | $1.58 |
| Paid from, on Premium | pool | credit | pool | credit | credit |
What this tells you:
- Cursor’s own models are effectively “free” until the pool runs out. Grok 4.7 and Composer 2.5 do not touch your credit while you have pool left. Their prices only apply after that.
- Claude costs money from the first request. In this example, $30 of credit buys about 38 sessions on Sonnet 5.5 or about 19 on Opus 5.5. The $30 Premium subscription itself covers none of them.
- Caching changes the ranking. Sonnet 5.5 has a cheaper cache read than Grok 4.7, so on long sessions it comes out cheaper even though its output costs more.
Legacy plans: what about Cursor Pro at $20?
Pro, Pro+, Ultra, Hobby and the old Teams seats keep working for existing customers, but new customers cannot buy them. According to the legacy plans page, moving to The X Subscription is optional, and once you switch you cannot go back.
| Legacy price | Closest new plan | |
|---|---|---|
| Hobby | $0 | Free |
| Pro | $20/mo | Premium ($30) |
| Pro+ | $60/mo | Super ($100) |
| Ultra | $200/mo | Ultra ($200) |
| Teams Standard | $40 per user | Premium seat ($30) |
| Teams Premium | $120 per user | Super seat ($100) |
The difference is in the pool. Legacy plans have two monthly pools: one for Cursor models and one for third-party models. The new plans have a single weekly pool, at least twice as large for Cursor models, but with no third-party allowance. If you mostly use Claude on Pro, moving to Premium means paying separately for every Claude token. If you mostly use Composer and Grok, you get more for the extra $10.
Which Cursor plan should you choose?
In my view the order is simple, and a bigger plan only pays off once you actually run out of pool.
- You want to see whether Cursor suits you: Free. If the limit gets in your way after a week, move up.
- You are learning and use the agent a few times a week: Premium at $30. Plus is cheaper, but it has no Auto, no third-party models and no cloud agents, and most tutorials rely on those.
- You want to work with Claude: Premium plus a top-up with a monthly additional usage limit. Set the limit before you turn on Auto Top-up.
- You spend several hours a day with the agent: Super. Cursor itself recommends it for daily agent users.
- You are on Pro and happy with it: change nothing. First work out how much Claude usage you get from the Other Models pool, because after switching you would pay for it separately.
I compare the terminal coding agents in Claude Code, Codex CLI and Gemini CLI: how they differ in practice, and I collect more on these tools in the coding agents hub. If you cancel, do it at least 24 hours before renewal to avoid another charge.
Once a week my newsletter covers what I check while building with AI: pricing changes in the tools, decisions and what they cost, plus one way of working I have checked myself.
Build with AI. Step by step.
I do not copy headlines. Once a week I write about what I test myself while building Frinter, the warehouse agent and sites in 1:1 workshops: what worked, what did not and why.
- Behind the scenes of my projects
- Decisions, mistakes and their cost
- One proven way of working
Last checked: November 21, 2026. Cursor notes that prices may change, so compare them with the pricing page linked in the sources below before you buy.

