Plan limits

How much Grok 4.7 usage do you get on Cursor?

Cursor gives Grok 4.7 no message count. It draws on the Cursor Models pool, which Pro, Pro Plus, Ultra and Start all include, at $2 per million input tokens and $6 per million output, and the pool resets monthly with your billing cycle.

Checked against the lab's own pages on September 29, 2026. Limits change often, so the date matters.

Grok 4.7 on Cursor doesn't come with a message count. It draws from the Cursor Models pool, a monthly allowance that's tracked separately from the pool third-party models like Claude and GPT use.

So running low on Opus doesn't eat into your Grok allowance, and the reverse holds too (the two pools are counted apart on your Spending tab).

Pool

Cursor Models, with Grok 4.6, Grok 4.5 and Composer 2.5

Standard

$2/M input tokens, $0.50/M cached input, $6/M output

Fast

$4/M input, $1/M cached, $12/M output

Context

256k standard window and 500k long context

Resets

monthly with your billing cycle

Which plans get it

Cursor's usage page says most plans carry two pools. Each plan it names carries the one Grok 4.7 uses.

On the Start plan (it's India only), Grok 4.7 is fixed at medium effort and standard speed. On Pro and higher, Fast is the default speed tier, and Fast costs twice the standard rate per token. So the default setting burns the pool quicker than the cheapest one does.

What makes a request expensive

Input size, mostly. Once input goes past 256k tokens, standard requests bill at 2x and Fast requests at 3x the standard rates, up to the 500k window. Cursor says long-context billing looks at input length only, so a request at or under 256k tokens bills at the normal rate however long the answer is.

Effort matters too. The model runs at four effort levels, xhigh, high, medium and low, with high as the default. Cursor's docs say harder tasks spend more time thinking at the higher settings.

What happens when the pool runs out

You'll get a notification in the editor. From there you can turn on on-demand usage, which is pay-as-you-go at the same API rates, or move to a higher plan. Usage resets monthly with your billing cycle and doesn't roll over, and the reset date is on the Spending tab of your dashboard.

One catch: Auto and subagents can spend the other pool. Cursor's help page says a subagent can run a named third-party model while the picker still shows Grok, and those requests bill Other Models.

Cursor doesn't print how many dollars or tokens the Cursor Models pool holds on any plan, so there's no fixed count of Grok 4.7 prompts to quote.