Cursor Teams Standard costs $40 per user per month, exactly double the $20 Pro plan, and adds centralized billing and administration, SAML/OIDC SSO, team-wide privacy mode, usage analytics, Bugbot agentic code review, cloud agents with shared team context, and a team marketplace for internal rules, skills, and plugins. It does not add a hard spend ceiling, per-repository cost attribution, coverage of any agent other than Cursor, or pooled usage, which is Enterprise-only. Buy it for the control surface and the single invoice, and plan the seat lifecycle before the first renewal.
- $40 per user per month, exactly double Pro. The model access is the same.
- The real product is administration: one invoice, SSO, verifiable privacy mode, analytics.
- No hard spend ceiling. Past the allowance, usage continues on demand and bills you.
- Analytics answer who used Cursor, not what a repository or a project cost.
- Pooled usage is Enterprise-only, so a mixed team over-buys on Teams by design.
- The largest avoidable cost is assigned seats nobody opens. Audit before renewal, not after.
The honest version of the decision
Cursor is the best AI editor in the category and has been for two years. If your engineers read and edit code all day, tab completion alone justifies the individual plan, and nothing in this page argues otherwise. The question here is narrower: whether the organization should buy seats centrally at $40 instead of letting people expense Pro at $20.
| If your binding constraint is | The answer is |
|---|---|
| Finance will not accept fifty expense reports | Buy Teams |
| Security requires verifiable privacy mode across the org | Buy Teams |
| IT requires SSO on every developer tool | Buy Teams |
| You want to know who is actually using it | Buy Teams |
| You want the bill to go down | Teams will not do that |
| You need a hard monthly ceiling | No Cursor tier provides one |
| You need cost split by repository or project | Cursor does not report that axis |
| Half the team uses Claude Code or Codex instead | A single-vendor seat is the wrong unit |
What the admin surface actually gives you
Cursor lists seven additions on Teams Standard. Taken one at a time, with what each is genuinely good for:
- Centralized team billing and administration. One invoice, one place to add and remove people. This is the item that most often carries the whole purchase, and it is worth more than it sounds if your finance team currently reconciles individual card charges.
- SAML/OIDC SSO. Developers sign in with the company identity provider rather than a personal account. Note that this sits on Teams, not Enterprise, which is unusual and in your favor.
- Team-wide privacy mode. Enforced centrally instead of trusted per person. A privacy setting that each developer can turn off is not a control you can put in front of a security review; this is.
- Usage analytics. Per-member activity, which is the input to seat reclamation and to the question of whether the rollout worked.
- Agentic code reviews with Bugbot. Automated review comments on pull requests. Judge this on your own repository during the trial rather than on the description.
- Cloud agents and automations with shared team context. Background agent runs that inherit team rules rather than one person’s local configuration.
- Team marketplace for internal rules, skills, and plugins. A distribution mechanism for conventions. Useful in proportion to how much your team has actually written down.
Cursor also exposes an admin API, which is what makes the analytics operationally useful rather than merely visible. Pull member activity on a schedule, keep the history, and you can answer the renewal question with evidence. Tracking Cursor usage covers the mechanics.
The four things it will not do
None of these are defects. They are scope boundaries, and every one of them has been mistaken for a bug by a team that assumed a paid organization tier implied them.
It will not cap your spend
When the included allowance is exhausted, Cursor does not degrade or stop. It continues on demand at the same API rates and adds the usage to your bill. Cursor documents this as a feature, and for a team under deadline it is one. It is also the reason a Cursor invoice can exceed the sticker price with nothing having gone wrong. The only levers are the routing default and, on Enterprise, model access controls.
It will not tell you what a project cost
Analytics are organized by member. That answers seat reclamation and adoption. It does not answer which repository, service, or client consumed the spend, which is the axis finance asks about when the number gets large enough to allocate. That is a general problem across seat-based AI tools and is covered in the cost allocation guide.
It will not pool usage
Pooled usage is listed under Enterprise. On Teams, allowances attach to seats, so a team with three heavy users and twenty light ones pays overage on the heavy end while the light end leaves allowance unspent. On a concentrated distribution this is the largest hidden cost of staying on Teams.
It will not cover your other agents
A Cursor seat covers Cursor. If half the team runs Claude Code or Codex, and most teams now run at least two agents, the seat is not the unit that matches the spend. You end up with one governed tool and an ungoverned remainder, which is a worse position than governing none of them because it looks solved.
A rollout order that survives the first renewal
The failure mode is buying seats for a headcount and discovering at renewal that a third of them were never opened. The order below is designed so that the seat count is derived from evidence rather than from an org chart.
Trial on individual plans first
Let twenty people expense Pro at $20 for a month. This costs $400 and produces the two numbers you need: how many people actually adopt it daily, and how much third-party model traffic they generate.
Count daily actives, not signups
The seat count to buy is the daily-active count plus a small margin, not the number of engineers. A seat assigned to someone who opens Cursor twice a month is a $480 annual donation.
Price the token surcharge
Multiply the measured third-party token volume by $0.25 per million, which is the Cursor Token Rate that applies on Teams and Enterprise and not on individual plans. Add it to the seat total before comparing against the trial cost.
Set the routing default deliberately
Auto Cost and Cursor first-party models are exempt from the surcharge and have their own generous allowance. Deciding the default centrally, and writing down when a frontier model is warranted, is the single largest lever on the eventual bill.
Wire the admin API into whatever you already use
Pull member activity on a schedule into your existing reporting. If the only place usage lives is a dashboard someone has to remember to open, the renewal conversation will be an argument about impressions.
Audit seats before the renewal boundary, not after
Removed seats generally bill to the end of the cycle. Reclamation is worth real money and only if it happens on the right side of the date.
Revisit Enterprise once the distribution is visible
If a month of data shows the top decile driving most of the traffic, pooled usage is the conversation to have, and you will have the evidence to have it well.
For a first pass at the arithmetic before any of this, the AI team cost calculator takes a headcount and a usage mix and returns per-seat and pooled totals side by side.
What a team-shaped alternative looks like
The reason to describe Continuum here is not that it replaces Cursor. It does not have tab completion and is not an editor. It is that the two products draw the org boundary in different places, and knowing where the other boundary sits is useful even if you keep Cursor.
A Cursor seat governs one tool completely. A Continuum seat governs the model spend across whichever agents your engineers actually run, and leaves the tools alone. If your team is single-vendor, the first shape is simpler and better. If your team has quietly become multi-vendor, which most have, the second shape matches the spend.
| Cursor Teams | Continuum org | |
|---|---|---|
| Seat price | $40/user/mo | $25 per seat per month |
| Billed on | Granted seats | Live member count |
| Covers | Cursor | Claude Code, Codex, Cursor CLI, Gemini, Grok, OpenCode |
| Editor and autocomplete | Yes, best in category | No, it is a workbench not an editor |
| SSO | Yes | Enterprise identity provider login |
| Hard spend ceiling | No | Weekly caps with approvals |
| Spend by repository | No | Yes |
| Client application price | $40 a seat | Free |
The full price breakdown for Cursor above the individual tiers is in the Cursor Enterprise pricing guide, and the side-by-side is on the Cursor comparison page. If you are evaluating whether to move rather than to add, the migration guide states plainly what you would lose.
Questions people ask
How much does Cursor cost for a team?
Teams Standard is $40 per user per month. Teams Premium is described as 5x Standard agent limits with no published price, and Enterprise is custom. There is no published volume discount on the pricing page.
Is Cursor Teams cheaper than everyone buying Pro?
No. It is exactly double: $40 against $20. The savings are administrative rather than financial. If cost reduction is the goal, a team plan is the wrong instrument.
Does Cursor Teams include SSO?
Yes, SAML/OIDC SSO is included on Teams Standard. It is not held back for Enterprise, which is worth knowing because it is the requirement most often used to justify an enterprise tier elsewhere.
Can I see which developers use Cursor the most?
Yes. Teams includes usage analytics by member, and there is an admin API so you can pull the same data on a schedule into your own reporting. Per-member activity is the input to seat reclamation.
Can I set a spending limit on Cursor Teams?
Not a hard one. When the included allowance is exhausted, requests continue on demand at API rates rather than stopping. The practical controls are the routing default and, on Enterprise, model access controls.
Does Cursor show cost per repository or project?
No. Analytics are organized by member. If you need spend attributed to a repository, service, or client, that axis has to come from somewhere else.
How many seats should we buy?
The daily-active count from a trial month, plus a small margin. Buying against headcount is how teams end up paying for seats nobody opened, which is the largest avoidable cost in any seat-based AI tool.
What if half the team uses Claude Code or Codex?
Then a Cursor seat governs a fraction of the spend. That is not an argument against Cursor, which most teams should keep. It is an argument for measuring total agent spend before deciding that a single-vendor seat is the governance unit.
Sources
Every figure above was read from these pages on August 2026. Vendors reprice without notice; if you find a stale number, tell us.
- Cursor pricing
- Cursor documentation
- Cursor admin API Per-member usage data behind the Teams analytics view.