Cursor pricing explained (2026): the real cost
verified 2026-05-31On Pro ($20/mo) you get $20 of included frontier-model usage. Token rates run roughly $1.25 per million input tokens and $6 per million output tokens for a model like Claude Sonnet, so $20 buys on the order of 200 to 300 premium agent requests per month depending on prompt and context size. Auto mode (Composer 2.5 at about $0.50 in / $2.50 out per million tokens) draws from a separate, much larger pool and is effectively unlimited for routine work, so it does not deplete the $20 frontier pool. Past the included amount, extra usage is billed in arrears at the same per-token API rates, so heavy frontier-model users can spend well above $20. Prices verified 2026-05-31 against cursor.com. prices change often, verify live
The tiers
Monthly and annual. The unit you pay against, per plan.
| tier | monthly | annual | what's included | unit |
|---|---|---|---|---|
| Hobby | $0 | $0 | Limited Agent requests, limited Tab completions, free trial of Pro | per user |
| Pro | $20 | $16/mo (billed yearly) | $20 of included model usage/mo, unlimited Auto + Composer usage, frontier model access, unlimited Tab, MCP/skills/hooks, Cloud Agents, Bugbot (usage-based) | per user/mo |
| Pro+ | $60 | $48/mo (billed yearly) | Everything in Pro plus ~$70 of included model usage/mo (3x Agent limits) | per user/mo |
| Ultra | $200 | $160/mo (billed yearly) | Everything in Pro plus ~$400 of included model usage/mo (20x Agent limits), priority access to new features | per user/mo |
| Teams | $40verify live | ~$32/mo per user (billed yearly) | Everything in Pro plus $20 of included usage per user/mo, pooled usage, centralized billing, usage analytics, org privacy mode, RBAC, SAML/OIDC SSO | per user/mo |
// pools and per-unit rates are volatile · verified 2026-05-31 · cells flagged verify live link to the live vendor page
The $20/mo "Pro" price is not what you actually spend.
The $20/mo "Pro" price is not what you actually spend. Since the June 2025 repricing, the subscription only buys $20 of frontier-model usage; selecting Claude/GPT/Gemini burns that pool at raw provider API rates and then bills overage in arrears. Only "Auto" mode is unlimited. Teams seats include the same $20 of usage per user, so a team of intensive coders can rack up large on-demand bills well past the per-seat fee.
See live pricing on Cursor →Worth it if
- You want an AI pair-programmer that lives inside your editor and can also dispatch autonomous background agents to fix bugs and ship PRs while you keep coding.
- Your usage stays inside the included allowance.
- You value AI coding / autonomous PRs over raw lowest cost.
Look elsewhere if
- The $20/mo "Pro" price is not what you actually spend. Since the June 2025 repricing, the subscription only buys $20 of frontier-model usage; selecting Claude/GPT/Gemini burns that pool at raw provider API rates and then bills overage in arrears. Only "Auto" mode is unlimited. Teams seats include the same $20 of usage per user, so a team of intensive coders can rack up large on-demand bills well past the per-seat fee.
- You're highly cost-sensitive at scale.
- A cheaper tool in the comparison covers your use case.
Ready to try Cursor?
Start on the free tier (Yes, Hobby plan at $0, no card. Limited Agent requests and Tab completions), then size your plan with the numbers above.
Affiliate link · how we make money
Questions
How much does Cursor really cost?
On Pro ($20/mo) you get $20 of included frontier-model usage. Token rates run roughly $1.25 per million input tokens and $6 per million output tokens for a model like Claude Sonnet, so $20 buys on the order of 200 to 300 premium agent requests per month depending on prompt and context size. Auto mode (Composer 2.5 at about $0.50 in / $2.50 out per million tokens) draws from a separate, much larger pool and is effectively unlimited for routine work, so it does not deplete the $20 frontier pool. Past the included amount, extra usage is billed in arrears at the same per-token API rates, so heavy frontier-model users can spend well above $20.
Does Cursor have a free tier?
Free tier: Yes, Hobby plan at $0, no card. Limited Agent requests and Tab completions.
What's the catch with Cursor's pricing?
The $20/mo "Pro" price is not what you actually spend. Since the June 2025 repricing, the subscription only buys $20 of frontier-model usage; selecting Claude/GPT/Gemini burns that pool at raw provider API rates and then bills overage in arrears. Only "Auto" mode is unlimited. Teams seats include the same $20 of usage per user, so a team of intensive coders can rack up large on-demand bills well past the per-seat fee.
Is Cursor worth it?
Cursor is worth it if you specifically need AI coding / autonomous PRs. That's where it beats the alternatives. On Pro ($20/mo) you get $20 of included frontier-model usage. Token rates run roughly $1.25 per million input tokens and $6 per million output tokens for a model like Claude Sonnet, so $20 buys on the order of 200 to 300 premium agent requests per month depending on prompt and context size. Auto mode (Composer 2.5 at about $0.50 in / $2.50 out per million tokens) draws from a separate, much larger pool and is effectively unlimited for routine work, so it does not deplete the $20 frontier pool. Past the included amount, extra usage is billed in arrears at the same per-token API rates, so heavy frontier-model users can spend well above $20. If your usage is high-volume or budget is the priority, a cheaper tool in the comparison will do the same work for less.