Cursor Doubles Usage Limits on All Individual and Teams Plans

Cursor doubled the included usage pool that its own models draw from on individual and teams plans, and the increase is permanent. A Cursor staff member confirmed it in writing on Cursor's official community forum: "the general 2x included usage is a separate thing and isn't tied to that date," meaning July 21, 2026, and a later reply in the same thread states plainly, "The doubled pool stays" (Cursor forum). The same staff thread separately confirms Grok 4.5's placement in the first-party pool is "a permanent plan change, not a promo."

What expired on July 21 was something else: a 50 percent launch discount on Grok 4.5. The two were widely conflated, which is why Cursor staff wrote a clarification at all.

Key Takeaways

• The doubled included usage is permanent, in Cursor's own words "The doubled pool stays" (Cursor forum).
• July 21, 2026 ended only Grok 4.5's 50 percent launch discount, not the pool doubling. "The doubled pool stays."
• The 50 percent discount and the doubled Grok usage were the same thing, not two multipliers. They did not stack to 4x.
• Per-token prices did not change. What doubled is the size of the included pool at those same rates.
• The doubling covers the Cursor Models pool: Cursor Grok 4.5 and Composer 2.5. Auto is not in this pool -- Cursor's docs now route Auto through third-party models via Auto Balance and Auto Intelligence. Third-party models draw from a separate pool that was not doubled.
• Nothing was recalculated on July 21. Usage already consumed stayed measured at the price active at the time.
• Enterprise is "a separate case with a single unified pool," so check your model controls and spend limits rather than assuming the doubling applies.

Is the 2x Included Usage Permanent?

Yes. This is the question Cursor users asked most directly, and Cursor answered it.

The confusion started with an in-product notice telling users their plan "now includes 2x usage for all Cursor models: Composer 2.5 and Cursor Grok 4.5." At the same time, Grok 4.5's model description carried the line "50% off through July 21, 2026." Users reasonably assumed both were one promotion expiring on the same day, and asked whether the 2x pool would disappear on July 22.

A Cursor staff member answered in the forum thread on July 19, 2026:

Separately from the Grok discount, there was a general 2x increase in included usage for first-party models, shown as 2x Included Usage for Composer 2.5 and Cursor Grok 4.5. This is separate from the Grok discount and it's not tied to July 21.

And, in the same reply:

So July 21 only ends the Grok 4.5 launch discount. The general 2x included usage is a separate thing and isn't tied to that date.

A follow-up the next day, answering a user who worried their consumed usage would be re-tallied when the discount lapsed, put it in four words: "The doubled pool stays."

That is a stronger statement than the July 21 announcement itself provided. Cursor's public post said only that limits were doubled, with no duration either way (Cursor on X). The forum reply is where Cursor committed to permanence in writing.

2x Included Usage vs the Grok 4.5 50% Discount

These are two different changes with confusingly similar wording. The distinction is the whole story.

Cursor also corrected the arithmetic that users had built on top of the confusion. The natural reading was that a doubled pool plus a half-price rate produced a 4x effective boost on Grok 4.5. Cursor staff replied:

Quick math fix: 50% off and doubled usage for Grok are the same thing, not two separate multipliers. Half the per-token price means you can do about 2x as much work from the same pool, not 4x.

The practical consequence, stated by Cursor: after July 21, "the same Grok work will use your pool about 2x faster than during the promo." If you calibrated your Grok 4.5 workload during the discount window, your pool now drains at roughly twice the rate it did then, even though the pool itself is still the doubled one.

What Actually Doubled: Pool Size, Not Per-Token Price

A recurring objection in the thread was that Cursor's published prices looked identical before and after, so nothing could have doubled. Cursor staff addressed this directly:

The per-token rate for first-party models (Auto / Composer 2.5 / Cursor Grok 4.5) really didn't change, so the pricing page looks the same as before. That's expected. What did change is the size of the included pool. The amount of usage included in the plan before switching to usage-based billing doubled. So the rate is the same, but the included quota at that rate is now 2x bigger.

This is why no public Cursor page shows a 2x figure. Checked August 1, 2026, Cursor's pricing page describes the Individual plan only as including "Generous limits for Grok & Composer" and the words "2x", "double", and "promotion" appear nowhere on it. Cursor's models and pricing documentation likewise documents per-token rates and pool structure without stating a multiplier. Cursor staff acknowledged the gap, calling the "Generous included usage" wording "vague" and pointing users to the billing dashboard for the actual number: "the exact remaining amount for your plan is in the billing dashboard."

So there is no public page you can cite to a finance team. The doubling is visible only in your own dashboard.

How Cursor's Two Usage Pools Work

Understanding which pool doubled requires knowing there are two. Per Cursor's documentation and the staff explanation in the forum thread:

• The Cursor Models pool covers Cursor's own models: Cursor Grok 4.5 and Composer 2.5. This is the pool that doubled. It is not expressed as a dollar amount. Note a docs change since the staff replies quoted below: Cursor's models-and-pricing page now states Auto is not part of this pool -- Auto references third-party models through the Auto Balance and Auto Intelligence modes.
• The Other Models pool is a dollar-denominated quota for third-party frontier models, documented as $20 on Pro, $70 on Pro+, and $400 on Ultra. This pool was not doubled.

Cursor staff described the consumption order as the Cursor Models pool first, then the included third-party quota, then on-demand billing. While the first pool has capacity, Composer 2.5 and Grok 4.5 draw from it and do not touch the third-party quota (the staff reply at the time listed Auto here too; current docs route Auto through third-party models instead). Once it is consumed, further first-party usage "spills over into your included API quota at the same per-token rates." Third-party models draw from the third-party quota from the start. Only when both are exhausted does pay-as-you-go billing begin, "and only if you have it enabled - otherwise requests are blocked until the cycle resets."

Note that Cursor has renamed these pools over time. The forum replies call them the "First-Party Models" pool and the "API pool"; the current documentation calls them the Cursor Models pool and the Other Models pool. Same structure, different labels depending on which surface you are reading.

Which Plans Get the Doubled Pool

Cursor's July 21 announcement said all individual and teams plans (Cursor on X). Two clarifications since then:

Enterprise works differently. Cursor staff described it as "a separate case with a single unified pool without first and third-party split." Because the doubling applies specifically to the Cursor Models pool, and enterprise does not split its pool that way, do not assume a 2x allocation there.

Cursor Start, launched July 28, 2026, is Cursor Models only. The India-only Start plan sits below Pro at ₹649 per month. Cursor's documentation states that "Start does not include the Other Models pool, on-demand usage, Bugbot, Auto, Automations, or the Cursor SDK," and the plan comparison table lists Start at $0 for third-party model usage. Start therefore lives entirely inside the Cursor Models pool, which is the pool that was doubled, but Cursor publishes no separate figure for Start's allocation.

How This Compares to the Rest of the Market

Vendors are competing on usage headroom, not just model quality, and as of August 2026 two of the largest lifts are running at the same time. Anthropic's +50% Claude Code weekly-limits promotion is still live: Anthropic's support article, re-checked August 1, 2026, states it runs through August 19, 2026 at 11:59 PM PT, having been extended from its original July 13 date.

The difference that matters is shape rather than size. Anthropic's lift is explicitly time-boxed with a published expiry you can plan around and a support article you can point at. Cursor's is open-ended and confirmed permanent, but only in a forum reply, with no public pricing page or documentation stating the multiplier.

How to Use This Update

If you throttled Composer or Grok usage to stay inside your allocation, the ceiling you tuned for is twice as high and staying that way, so remove the self-imposed rationing and re-test your real workflow.

If you benchmarked your Grok 4.5 spend between July 8 and July 21, re-measure it. Your pool is unchanged but Grok 4.5 now consumes it about twice as fast as it did during the discount, so a workload that fit comfortably in the promo window may not fit now.

If you routed overflow work to a bring-your-own-key model to conserve pool usage, re-run the math: a permanently doubled first-party pool may make Cursor's own models the cheaper path again, particularly since first-party usage spills into your third-party quota only after the first pool is empty.

And if you need the actual number rather than a multiplier, check the billing dashboard. Cursor does not publish per-plan pool sizes on any public page.

Sources

• Cursor staff reply, "Clarification needed: Is the '2x Included Usage' temporary, and how does it relate to the 50% discount?", Cursor community forum, July 17 to July 20, 2026. Every quotation about permanence, the discount distinction, the 4x math correction, pool mechanics, consumption order, and enterprise comes from replies in this thread by the account deanrie, which the forum's own API reports as staff with the group CursorStaff. Staff status verified August 1, 2026.
• Cursor, "Models and Pricing" documentation. Cursor Models pool composition, the Other Models pool at $20 / $70 / $400, on-demand billing, and the Start plan exclusions. Checked August 1, 2026.
• Cursor pricing page. Individual plan wording "Generous limits for Grok & Composer"; no multiplier, promotional window, or expiry stated anywhere on the page. Checked August 1, 2026.
• Cursor, "Grok 4.5" blog post, July 8, 2026. Grok 4.5 at $2 per million input tokens and $6 per million output tokens, Fast variant at $4 and $18, and the "double usage for the first week" launch wording. Re-checked August 1, 2026.
• Anthropic support article, titled "Claude Code May-August 2026 weekly limits promotion". Re-checked August 1, 2026.
• Cursor on X, July 21, 2026, the original announcement. X returns HTTP 402 to automated fetches, so this citation is preserved as published and could not be re-verified this run. Every claim it carries is independently sourced above.