Cursor Please Enable Usage-Based Pricing: What It Means Now

Cursor’s Please Enable Usage-Based Pricing prompt means your included usage is spent, or the model you picked isn’t covered, and Cursor wants permission to bill beyond your plan. What “yes” turns on depends on your plan type: legacy plans (Pro, Pro+, Ultra, Teams) switch on on-demand usage, while The X Subscription plans (Plus, Premium, Super, Ultra) use prepaid top-ups instead.

So nothing is broken: the message is Cursor asking whether you want to start paying beyond your subscription. Before you click yes, you need three answers: which plan type you’re on, what happens to your card once extra usage is on, and whether a cheaper model or a hard cap gets you through the cycle instead.

Which “Please enable usage-based pricing” message you saw

The exact wording tells you which era of Cursor billing you’re dealing with. On June 16, 2025, Cursor’s pricing announcement moved Pro from 500 requests a month to usage billed at model API prices, and the prompt has been reworded several times since. These are the variants users have reported on the Cursor Forum, plus the two limit states Cursor documents itself:

If you’re seeing that wording, you’re on a legacy request-based plan or an old client. Cursor’s docs don’t confirm that the exact “Please enable usage-based pricing” string appears on The X Subscription plans at all, so check your plan type before following any fix.

How to tell which plan you’re on: legacy plans are managed at cursor.com/dashboard and have a Spending tab. The X Subscription is managed at grok.com/dashboard, with billing at grok.com/dashboard/billing. Cursor says new customers start on The X Subscription, while existing customers can stay on legacy plans, which keep working.

What usage-based pricing means on Cursor now

According to Cursor’s account pricing docs, Cursor is now sold as part of The X Subscription: Free, Plus at $8, Premium at $30 (4x Plus usage), Super at $100 (15x) and Ultra at $200 (50x). The legacy plans (Hobby, Pro, Pro+, Ultra and Teams; Cursor no longer lists their prices) work differently enough that the same click means two different things:

Two details matter for the bill. Top-ups add prepaid credit to your subscription, and Cursor’s pricing docs don’t say whether that credit expires, so buy only what you expect to use. And on legacy plans, on-demand usage is something you enable on the Spending tab, on top of the two monthly pools.

Why your Cursor bill is higher than your plan price

The root cause goes back to June 16, 2025, when Pro stopped meaning 500 requests and started meaning $20 of included usage at API prices, with “unlimited” applying to Auto only. The rollout confused enough people that Cursor’s July 4, 2025 post, “Clarifying our pricing,” took “full responsibility” and offered refunds for unexpected usage between June 16 and July 4.

The mechanism hasn’t changed since: cost follows tokens times the model’s price, and in Cursor’s words the hardest requests cost “an order of magnitude more” than simple ones. Today the usual reasons a bill lands above the plan price are:

• Third-party models are billed separately. On The X Subscription, every Claude, GPT or Gemini request is paid from prepaid credit. On legacy plans, they drain the Other Models pool, then go to on-demand.
• Auto isn’t free. Router modes other than Cost bill at the list price of whichever model a request is routed to.
• Teams pay a token rate. Teams and Enterprise pay a Cursor Token Rate of $0.25 per 1M tokens on eligible third-party requests. Individual plans don’t, and Cost mode and Cursor models are exempt.

How to enable usage-based pricing and set a spend limit

Everything lives on the web dashboard, not in the editor settings. Use the block that matches your plan type; the tabs don’t exist on the other one.

Legacy plans: on-demand usage

1. Go to cursor.com/dashboard and open the Spending tab (cursor.com/dashboard/spending).
2. Turn on on-demand usage.
3. If the tab offers a spend limit, set one now. Cursor’s usage-limits help page doesn’t document those settings, so check what the tab shows before relying on it.
4. Check usage for both pools, remaining allowance and the reset date on the same Spending tab.

The X Subscription: top-ups and a monthly limit

1. Go to grok.com/dashboard/billing.
2. Under Additional usage credits, set a monthly additional usage limit.
3. Buy a top-up, which adds prepaid credit. Auto Top-up, which refills when your balance drops below a threshold you set, stays off unless you enable it.
4. Check usage on the Usage page at grok.com/dashboard.

Cursor’s pricing docs say the monthly additional usage limit caps how much paid additional usage you can consume each month.

At the limit, Cursor’s usage-limits help page lists the options: on a legacy plan, enable on-demand usage or upgrade; on The X Subscription, keep working from prepaid credit, wait for the weekly reset, or upgrade.

What each model costs once your included pool is gone

Once included usage runs out, or for any third-party model on The X Subscription, the per-model list price is the number that sets your bill. These come from Cursor’s models page:

Prices checked November 1, 2026 on cursor.com/docs/models. Max Mode isn’t in this table because Cursor offers it only on legacy request-based plans, at API rate plus 20%.

To put the spread in dollars, here’s plain arithmetic on those list prices, ignoring cache discounts. A chunk of work using 1M input tokens and 100k output tokens costs:

• $0.15 on Claude Haiku 5.5
• $0.75 on Composer 2.5
• $1.125 on Gemini 3.8 Flash
• $3.00 on Claude Sonnet 5.5
• $6.00 on Claude Opus 5.5
• $15.00 on Claude Fable 5.1

Fable 5.1 costs 20x what Composer 2.5 does for the same token count, and 100x Haiku 5.5. Model choice is the single biggest lever you control once extra usage is on. Cursor doesn’t publish the dollar size of any included pool, only the 4x/15x/50x multiples between X Subscription tiers, so there’s no official way to say how many of these chunks a plan covers.

Why switching models sometimes clears the message

On legacy plans the two pools empty separately. If Other Models runs dry, Composer 2.5 or Grok may still be served from the Cursor Models pool, and the limit message goes away without spending anything. On The X Subscription the logic is similar: Composer and Grok draw from the weekly included pool, while every third-party request costs credit, so moving to a Cursor model is the free option there too.

Auto is no longer a free escape hatch

Older pricing explainers still describe Auto as unlimited or flat-rate. That’s no longer how Cursor documents it. Auto was relaunched as the Cursor Router (documented on Cursor’s models page), with three modes: Cost, Balance and Intelligence. The old Auto is now called Cost, and Cost mode bills as a Cursor-model request from the included pool. Individual plans get Cost mode only; Balance and Intelligence, described for Teams and Enterprise with the Router on, bill at the list price of whatever model each request lands on. Our Cursor Auto mode cost guide goes through when each mode is worth it.

Should you enable it? Five options by situation

Enabling extra usage is one option among five, and for many people it isn’t the first one to try.

Switch models first. This costs nothing if Cursor-model usage still has room. Composer 2.5 at $0.50 input / $2.50 output per 1M is the cheaper of the two Cursor models. Among third-party models, Claude Haiku 5.5 and GPT-5.6 Luna are the cheapest per token, but on The X Subscription they’re still paid from credit. The trade-off is quality on hard multi-file refactors, where you may end up re-prompting.

Auto Cost is a reasonable default. It draws from the Cursor-model pool instead of credit, so for individual plans it’s the predictable choice. If a task needs a particular third-party model, pick that model yourself and accept its price.

Waiting is underrated. X Subscription usage resets weekly; legacy pools reset monthly with your billing cycle, and the date is on your usage page. Unused usage doesn’t roll over. Waiting isn’t free usage, though: until the reset, Cursor’s options are adding usage or upgrading. If you’re only a few days out (or hours, on a weekly plan) and not on a deadline, that pause costs nothing extra.

A small hard limit is the safe way to say yes. On The X Subscription, set the monthly additional usage limit to an amount you’d be fine losing, before the next request; on a legacy plan, set a cap on the Spending tab if it offers one. You keep working and the worst case is roughly known. Who shouldn’t do this: anyone tempted to leave extra usage uncapped just to make the prompt go away.

Upgrade only if the overage repeats. On legacy plans that means a higher legacy tier (Pro+ or Ultra; Cursor no longer lists their prices). On The X Subscription, Premium ($30) is the first tier that can use third-party models at all, and Super and Ultra raise Cursor-model usage to 15x and 50x Plus. If you blew past the limit once during a big refactor, don’t upgrade over a one-off.

Legacy wording? Check your plan. If your message says “slow pool” or mentions o1, it comes from the old request-based system. Confirm your plan in the dashboard before following any fix.

For context on how this compares with a competitor’s metering, see how GitHub Copilot’s AI credits billing works.

Set a cap before you click Enable

If Cursor shows you Please Enable Usage-Based Pricing or a usage-limit notice and you need to keep coding, try Composer 2.5 or Grok first; on either plan type they draw from included usage, not your card. If the task needs a third-party model, set the cap before you pay anything else: on a legacy plan, open cursor.com/dashboard/spending, enable on-demand and set a cap if the tab offers one; on The X Subscription, set the monthly additional usage limit at grok.com/dashboard/billing before buying a top-up. Don’t leave extra usage uncapped unless someone else is watching the invoices.