Cursor Usage Limit Explained: What Happens When You Run Out of Included Usage

Claude, Codex, Grok, and Cursor expose usage in different ways. Claude uses a rolling five-hour session window, with weekly limits on paid plans; Codex limits vary by plan; paid Grok plans use one shared weekly pool. Cursor's current individual plans instead show two usage pools that reset with the monthly billing cycle. If you are asking "how much Cursor usage do I have left?", the key is which pool you used and whether on-demand billing is enabled.

Cursor doesn't count requests anymore

Older Cursor plans used request counts; those rules still apply to some legacy plans. Current Pro, Pro Plus, and Ultra plans show two separate monthly usage pools:

• Cursor Models pool — usage for Cursor's models, including Grok 4.7, Grok 4.6, Grok 4.5, and Composer 2.5.
• Other Models pool — usage for third-party models such as Claude, GPT, and Gemini, charged at the selected model's listed API rate.

Both pools and their remaining allowance are visible in Cursor's Spending dashboard. Model choice and task size affect consumption, so an included amount does not translate to a fixed number of requests. Cursor documents the pool details in its usage guide.

Plans and what they include

These are the individual list prices shown on Cursor's pricing page as checked on 2026-10-07. Team and enterprise billing has separate terms. More importantly, a plan price is not a request count: included usage depends on the model and what you ask it to do. Check the official pricing page and your Spending tab before choosing a tier.

What actually happens when you hit the limit

Once a Cursor included pool is spent, the next step depends on the on-demand setting. Cursor's usage guide says you can enable on-demand usage for additional requests, billed by model usage, or upgrade your plan. If on-demand is off, usage pauses until the pool resets on the next billing cycle.

1. Enable on-demand usage — keep working and pay for additional model usage under Cursor's current billing terms.
2. Upgrade your plan — move to a higher tier for more included usage instead of paying overage rates.

Set a monthly spending cap before enabling on-demand usage. Cursor says included usage resets monthly with the billing cycle and unused usage does not roll over; the Spending tab shows remaining allowance and on-demand charges.

One extra line item to know about: Cursor's Models & Pricing documentation says third-party model requests on Teams and Enterprise plans carry a Cursor Token Rate of $0.25 per million tokens on top of the model's API price. Cursor's own models are exempt.

Why this is a different mental model, not just different numbers

The comparison is not simply "time limit versus money limit." Anthropic describes a rolling five-hour Claude window plus weekly limits on paid plans; OpenAI says Codex limits vary by plan and shows the current reset in the account usage view; xAI describes a shared weekly pool for paid Grok plans. Cursor's current individual plans use monthly pools, with optional on-demand billing. Each provider's dashboard is the source for the actual allowance on your account.

The meshcode angle

meshcode is a native desktop app for macOS and Windows with multiple agent panes. You can connect your existing Cursor-compatible model access or Claude Code/Codex CLI; the provider's billing and limits still apply. meshcode also offers a prepaid model option, whose cost depends on the model and task. Compare the actual usage you see before deciding it will be cheaper. For provider-specific limits, see Claude Code, Codex CLI, and Grok.

Sources and currentness

Checked on 2026-10-07: Cursor plans, Cursor usage limits, and Cursor Models & Pricing. For comparisons, see Anthropic's usage information, OpenAI's Codex usage guide, and xAI's Grok FAQ. Plan details change; confirm them in your account before relying on a quoted limit.

👉 Download meshcode — Mac, Windows