All insights

Coding agents

Cursor Cloud Agents Pricing: Plans, Usage Costs, and API Limits

Cursor’s US pricing lists Premium at $30/month with Cloud Agents included. See plan prices, shared usage, extra model charges, and what the spending limit does—and does not—cap.

Illustration for Cursor cloud agents pricing

The short answer

Cursor includes Cloud Agents with Premium and higher plans; the US pricing page does not list a separate Cloud Agents subscription price. Premium is $30/month or $300/year, before applicable taxes. Its weekly usage pool is shared across Cursor and Grok products, not reserved for Cloud Agents. Third-party models are billed separately at API list prices from prepaid credit. There is no published flat price per Cloud Agent run in the pricing information cited here. (Cursor Models & Pricing)

The practical budget is the plan fee plus any additional usage. Cursor publishes relative pool sizes, not a guaranteed number of Cloud Agent runs. Costs depend on model choice and usage. If by “production model API” you mean programmatically launching agents, Cursor documents a Cloud Agents API for agent workflows; that is different from a general model-inference API.

Cursor plan prices and Cloud Agents access

The following are prices shown on Cursor’s US-facing pricing pages. Prices exclude applicable taxes. Existing subscribers on legacy Cursor plans have different terms; check the plan attached to your account before comparing or switching. (Cursor Models & Pricing)

Plan Listed US price (USD) Cloud Agents Included usage
Free $0/month No Limited usage
Plus Not shown here No Base weekly usage for Cursor models
Premium $30/month; $300/year Yes 4× Plus usage
Super $100/month; $1,000/year Yes 15× Plus usage; priority model access at peak times
Ultra $200/month; $2,000/year Yes 50× Plus usage; earliest access to new products and models
Teams $30, $100, or $200 per paid seat/month Yes Usage matches the seat’s tier
Enterprise Custom Yes Pooled usage; contact Cursor for terms

Cursor lists annual prices for Plus, Premium, Super, and Ultra; the cited pricing information does not list annual prices for Teams. Premium is the lowest listed new-customer plan that includes Cloud Agents. Its $300 annual price is $60 less than 12 monthly payments at $30, before tax. Teams seats can use different tiers, so the total depends on the seat mix. (Cursor Team Pricing)

The plan multipliers describe included usage—not a promised number of cloud runs or a per-run discount. Cursor says the weekly pool for Cursor models is shared across products, including Cloud Agents. Heavy agent use can leave less included usage for other Cursor or Grok activity. (Cursor Models & Pricing)

What is included—and what can cost extra

Cursor-model usage: Paid plans include a weekly pool for Cursor models. It is shared across Cursor and Grok products, refreshes weekly, and unused included usage does not roll over. Models draw from the pool at different rates. The listed multipliers are relative to Plus; they do not give an absolute dollar allowance or a fixed number of Cloud Agent tasks. (Cursor Models & Pricing)

Third-party models: These are not included in the Cursor-model pool. On Premium and above, Cursor says third-party model usage is charged at API list price from prepaid credit. The plan fee alone therefore does not cover workflows that use those models. (Cursor Models & Pricing)

Teams and the Cursor Token Rate: Teams and Enterprise can also incur a Cursor Token Rate of $0.25 per million tokens on eligible third-party model requests, in addition to the model’s API price. Cursor documents exemptions for Cost-mode requests and for cache-read prices strictly below $0.10 per million tokens across every pricing tier. The cache-read exemption also applies when using a personal API key. Individual plans do not pay this rate. (Cursor Token Rate)

Automations: Cursor says automations create Cloud Agents and are billed based on Cloud Agent usage. For team automations, the selected run identity determines whether usage is billed to the person who created the automation or to the team’s usage pool. Cursor also says automations use each model’s maximum supported context window, which can affect usage. (Cursor automation documentation)

After the included pool is used

On The X Subscription, Cursor says you can continue using prepaid credit, wait for the weekly pool to refresh, or upgrade. Prepaid credit does not reset with the weekly pool; Cursor says top-ups currently expire 365 days after purchase. Usage and the next reset date are shown in the dashboard. (Cursor Usage and limits)

Cursor also documents a monthly limit for paid additional usage. When reached, paid additional usage stops while included weekly usage continues. The cited spend-limit information does not establish when paid additional usage resumes or whether raising the limit restores it, so confirm the reset and resumption conditions for the account. The limit counts paid additional usage consumed, not prepaid credit purchased. Buying more prepaid credit does not count toward the limit, so the limit does not cap top-up purchases. (Cursor Spend limits)

A budget example without guessing run costs

For one subscriber on monthly Premium, the listed fixed fee is $30/month before tax. Cloud Agent use of Cursor models draws from the shared weekly pool. If that pool is exhausted, the user can wait for its reset, upgrade, or continue with prepaid credit at applicable rates. Third-party model use is billed separately at API list prices from prepaid credit; eligible Teams and Enterprise requests may also incur the Cursor Token Rate. (Cursor Models & Pricing)

That identifies the cost categories, but it does not produce a reliable per-run estimate. Cursor does not list a flat Cloud Agent run price, and the cited plan information gives relative pool sizes rather than an absolute task allowance. A workload estimate would need the model route, usage per run, remaining shared pool, and—where relevant—the third-party API price and Token Rate eligibility.

Can a production model API be used with Cloud Agents?

“Production model API” can mean two different workflows:

  1. Launching and managing agents through Cursor’s API: Cursor documents the Cloud Agents API for programmatically launching and managing agents and their environments. Cursor’s API overview describes agent workflows—workspace context, tools, commands, and edits—not standalone model inference or chat completions. The Cloud Agents API v1 is labeled public beta, and Cursor warns that APIs may change before general availability. Treat that status as an integration consideration if you need a stable production interface. (Cloud Agents API) · (Cursor APIs overview)
  2. Using your own model-provider API key for Cloud Agent inference: Cursor’s BYOK documentation describes API keys in Cursor settings, but the cited material does not establish that a key configured for local editor use is used by Cloud Agents. Confirm support and billing for the exact cloud-agent route before assuming editor BYOK behavior applies. (Bring your own API key)

Cloud Agents can also use MCP servers to access external tools and data sources, including APIs and third-party services. That is tool connectivity; it does not establish that an external model API can replace Cursor’s model routing. Budget any external service charges under that provider’s own terms. (Cloud Agents documentation)

What to verify before budgeting

  • Plan and billing period: Confirm the subscription tier, monthly or annual billing, applicable taxes, and whether an existing legacy plan has different terms.
  • Usage pool: Check remaining weekly usage and the next reset date. The pool is shared, not reserved for Cloud Agents, and the listed multipliers do not translate directly into a fixed run count.
  • Model route: Distinguish Cursor-model use from third-party model use. For Teams and Enterprise, check whether a third-party request is eligible for the Cursor Token Rate.
  • Additional-usage limit: Set a monthly limit if appropriate. It limits paid additional usage, not prepaid-credit purchases; check the account’s reset and resumption conditions.
  • Automation billing identity: For team automations, check whether the run is charged to an individual or the team pool.
  • API workflow: For programmatic agent launches, account for the Cloud Agents API’s public-beta status. If you mean external model inference through a provider key, confirm the specific Cloud Agent support and billing route first.

Bottom line: Premium is the lowest listed new-customer plan with Cloud Agents at $30/month or $300/year in the US-facing pricing, before tax. That buys access and a shared weekly Cursor-model pool—not a dedicated per-agent allowance. Budget separately for additional usage and third-party models; eligible Teams and Enterprise requests may also incur the $0.25-per-million-token Cursor Token Rate. Cursor does not publish a flat cost per Cloud Agent run in the pricing information cited here.

Sources

Cursor Pricing (US)

Models & Pricing | Cursor Docs

Team Pricing | Cursor Docs

Cursor Token Rate | Cursor Docs

Usage and limits | Cursor Docs

Spend limits | Cursor Docs

Automations | Cursor Docs

Cloud Agents API | Cursor Docs

Cursor APIs Overview | Cursor Docs

Bring your own API key | Cursor Docs

Cloud Agents | Cursor Docs

Contact us