What to do when Cursor hits its usage limit: 9 ways to keep coding in 2026

Disclosure: Enjoy is our product, and it runs Claude Code and Codex, not Cursor. We checked every price and limit below against Cursor’s, SpaceXSI’s, Anthropic’s, OpenAI’s and OpenCode’s own pages on November 3, 2026, and we say where another way is the better pick, including the free ones that don’t need Enjoy.

The ways to keep coding, compared

When your included usage runs out, Cursor’s usage page says “You’ll see a notification in the editor.” A banner quoted on its forum in July opened “You’re out of usage.” What you can do next depends on which of Cursor’s two kinds of plan you’re on:

• The X Subscription, where new customers have started since October 12: “every paid plan includes one weekly pool of usage for Cursor models (Grok 4.7, Grok 4.6, Grok 4.5, and Composer 2.5), shared across Cursor, Grok Bot, and the Grok apps.” Claude, GPT and Gemini aren’t in it.
• A legacy plan, such as Pro, Pro+ or Ultra: “included usage is split into two monthly pools,” Cursor Models and Other Models, the second for other labs’ models.

Cursor doesn’t publish how big any pool is. What to do depends mostly on which pool ran out and how far off the reset is.

Way	Cost, checked November 3, 2026	How soon you’re back	Best for
Wait for the reset	Free	Up to a week on The X Subscription; up to a month on a legacy plan	A weekly pool that resets soon
Switch to Grok or Composer	Free, while that pool has usage left	Right away	Other labs’ models that ran out first
Pay as you go	Each model’s rate: Composer 2.5 is $0.50 in and $2.50 out per million tokens, Claude Opus 5.5 $4 and $20	Right away	Finishing the task in front of you
Upgrade your plan	Premium $30 a month; Super $100; Ultra $200; legacy Pro+ $60	Right away	Running out most weeks
Your own API key	The provider’s API rates	Right away, in Cursor’s Chat and Agent	Other labs’ models, billed by the lab
Claude Code, inside Cursor	Claude Pro $20 a month, or $17 billed yearly; Max $100; Ultra $200	Right away, in a new conversation	A Claude plan you already pay for
Codex, inside Cursor	ChatGPT Plus $20 a month; Pro $100, $200 or $500	Right away, with your Cursor chats if you import them	A ChatGPT plan you already pay for
OpenCode, with free or local models	Free	Right away, in a new conversation	No bill and no plan to run out
Enjoy	A monthly subscription, free for 7 days, with your phone and other computers (pricing)	Right away, in a new conversation	Several labs’ agents, so the next limit doesn’t stop the work

Prices are each company’s US list prices, before tax. Waiting, switching models and OpenCode’s free models cost nothing extra; the other ways add to what you pay for Cursor.

Wait for the reset

Best for: a weekly pool that resets in a day or two.

Cursor’s usage page says where to look. On The X Subscription, Usage at grok.com/dashboard “shows how much of your weekly pool you’ve used, the next reset date, and your prepaid credit,” and “Your editor settings show the same remaining usage.” On a legacy plan, it’s the Spending tab at cursor.com/dashboard. The X Subscription’s pool “refreshes weekly,” and on a legacy plan “usage resets monthly with your billing cycle.”

Price: free.

Watch for: “Unused included usage does not roll over on either kind of plan,” and a legacy plan’s reset can be weeks away. Free’s usage is listed only as “Limited,” with no reset date on these pages. On The X Subscription, the pool “is shared across everything on your subscription: Agent in the editor, CLI, Cloud Agents, Grok Bot, and the Grok apps,” so if you run Grok Bot on Premium or above, it can spend your week while you aren’t coding. On November 1, many Grok Bot users saw their weekly usage jump to 100%, as our memo explains.

Switch to Grok or Composer

Best for: Claude, GPT or Gemini that ran out while Cursor’s own models still have usage.

The two kinds of usage run out separately. On The X Subscription, other labs’ models are paid from prepaid credit “from the first request,” and Cursor says “Grok and Composer draw from your included pool instead of prepaid credit.” On a legacy plan, other labs’ models draw on Other Models, and Grok and Composer on Cursor Models, so one pool can be empty while the other isn’t.

It also makes what’s left last. “Different models draw from the pool at different rates,” says Cursor’s models page, which lists Composer 2.5 at $0.50 per million input tokens and $2.50 per million output, against $2 and $6 for Grok 4.7. Leave Fast off: Composer 2.5 Fast is $3 and $15.

Price: free, from the pool you have left.

Watch for: it doesn’t help once your pool is empty. “After a seat uses up its included weekly usage, any further first-party usage is Additional Usage, paid from additional usage credits bought with top-ups,” says SpaceXSI. On legacy plans, don’t count on Auto either. In July, Cursor’s staff wrote that after 100%, “Auto and Composer 2.5 are meant to keep working,” and in August, they wrote “Auto isn’t unlimited,” and that to keep going “you’ll need to wait for your cycle to reset, upgrade your plan, or enable on-demand usage.” Plus runs at “fixed settings (Fast off, medium effort) without Auto.” And “Subagents can also run a named third-party model while the picker still shows Auto, Grok, or Composer,” which bills as third-party usage.

Pay as you go: prepaid credit or on-demand usage

Best for: finishing the task in front of you today, without changing plans.

On The X Subscription, “Top-ups add prepaid credit to your subscription,” per Cursor’s billing help. Buy credit under Additional usage credits at grok.com/dashboard/billing, or turn on Auto Top-up, which is “off until you turn it on.” On a legacy plan, on-demand usage “lets you keep working after an included pool runs out,” per its legacy plans page, and on Pro, Pro+ and Ultra it’s “Off until you turn it on under Spending.” Credit comes with a promise that “Requests are never downgraded in quality or speed,” and on-demand usage with “no change in quality or speed.”

Price: the model’s rate per million tokens, on Cursor’s models page: Composer 2.5 $0.50 in and $2.50 out, Grok 4.7 $2 and $6, Claude Sonnet 5.5 $2 and $10, Claude Opus 5.5 $4 and $20. On Teams and Enterprise, eligible third-party requests add a “Cursor Token Rate of $0.25 per million tokens.”

Watch for: set a cap before you start. On The X Subscription, a monthly additional usage limit stops paid usage “even if credit remains.” On a legacy plan, on-demand usage is billed “in arrears,” so the spend limit is what stands between you and a surprise. Either way, “Enforcement is not instant, so usage can briefly exceed your spend limit.” Third-party models are “Premium and above only.” On legacy Start and Hobby, on-demand usage is “Not available. AI features pause until the pool resets.” On a team, “Only admins can buy credit or change Auto Top-up.” And GPT may not last in Cursor: OpenAI has proposed to end its models there on November 12.

Upgrade your plan

Best for: running out most weeks, not once in a while.

Cursor’s pricing sets Plus, at $8 a month, as the base: Premium, at $30, has “4x more usage compared to Plus,” Super, at $100, 15 times, and Ultra, at $200, 50 times. By those multiples, a Plus-sized week costs $8 on Plus, $7.50 on Premium, about $6.67 on Super and $4 on Ultra. On The X Subscription, “Upgrades apply right away.” On a legacy plan you can “move between legacy plans, for example from Pro to Pro+, without switching,” and Pro+, at $60, includes “more of each pool than Pro.”

Price: Premium $30 a month, Super $100 and Ultra $200, or $300, $1,000 and $2,000 a year. On a legacy plan, Pro+ $60 and Ultra $200.

Watch for: a bigger plan buys a bigger share of a pool Cursor doesn’t size, and on The X Subscription, “Third-party models aren’t included” at any tier. Upgrading a legacy plan just after you run out may earn little credit: “Any credit for your previous plan is based on its unused included usage, not days remaining.” And moving from a legacy plan to The X Subscription is a bigger step. Each paid plan there “includes at least twice as much coding usage with Cursor models as the legacy plan it replaces,” but “It also drops the included third-party model allowance,” and “The switch is permanent. It can’t be undone, and Support can’t undo it for you.”

Your own API key

Best for: Claude, GPT or Gemini in Cursor’s own Chat and Agent, billed by the lab instead of through Cursor.

Add a key under Cursor Settings > Models. “On individual plans, this means unlimited AI messages at your own cost,” says Cursor’s help page, and those requests “don’t draw from your included usage or prepaid credit.” It takes keys from OpenAI, Anthropic, Google, Azure and AWS Bedrock.

Price: the lab’s API rates. Anthropic lists Claude Opus 5.5 at $4 per million input tokens and $20 per million output, and OpenAI lists GPT-6.1 Sol at $2 and $10.

Watch for: it needs Premium, Super or Ultra, or a legacy Pro, Pro+ or Ultra plan; Plus has no “third-party models, or your own API keys.” “Custom API keys only work with chat models,” and OpenAI’s help page adds that “Cursor doesn’t support ‘Bring your own key’ in Cursor Tab, Auto, Cloud or Background Agents, Automations, the Cursor CLI, or Cursor’s API and SDK.” Cursor’s “Zero Data Retention policy does not apply when you use your own API keys.” And a chat subscription isn’t an API key: “A ChatGPT subscription does not include OpenAI API usage.”

Claude Code, inside Cursor

Best for: a Claude plan you already pay for, with Cursor still your editor.

Claude Code’s extension has an “Install for Cursor” link, and “any paid Claude subscription (Pro, Max, Ultra, Team, or Enterprise) or a Claude Console account works, and no API key is required.” Anthropic’s help center says your plan “also covers Claude Code in supported IDEs, including VS Code, Cursor and other VS Code forks,” so it draws on Anthropic’s limits, not Cursor’s pool. A project’s instructions can carry over too: Cursor treats AGENTS.md “as an alternative to .cursor/rules,” and Claude Code “can read AGENTS.md as your project instructions.”

Price: Claude’s pricing has Pro at $20 a month, or $17 a month billed yearly, Max at $100 and Ultra at $200. The free plan leaves Claude Code out.

Watch for: Claude has limits too: “Pro and Max plans have a five-hour session limit and a weekly limit,” and “IDE usage counts toward the same usage limits shared across Claude and Claude Code.” Our Claude Code usage limit guide covers what to do then. It starts a new conversation, and “Some features are only available in the CLI.” Check where you got the extension: on October 29, Cursor’s marketplace served “a maliciously modified version” between about 2:27 a.m. and 1:35 p.m. Pacific, and our memo shows how to check a machine.

Codex, inside Cursor

Best for: a ChatGPT plan you already pay for, and OpenAI’s GPT-6 models, which Cursor doesn’t list.

OpenAI’s help page says the Codex extension “lets you use the Codex agent directly inside Cursor,” and that “It runs as a separate extension and does not depend on OpenAI models being available through Cursor’s own model picker.” Select Sign in with ChatGPT and it uses your plan’s limits. It’s also the one way here that can bring your work along: OpenAI’s import guide says the desktop app and Codex CLI can import from Cursor, and “Codex CLI imports up to 50 chats from the last 30 days.” It reads your instructions too: “Codex reads AGENTS.md files before doing any work.”

Price: ChatGPT Plus is $20 a month, and Pro $100, $200 or $500. Codex’s pricing lists the IDE extension from Plus; Free and Go get “GPT-6 Luna at Standard speed in the desktop app, subject to rollout.”

Watch for: OpenAI estimates 15 to 160 local messages every five hours with GPT-6.1 Sol on Plus, but “These estimates are not fixed message limits,” and “Weekly limits may also apply.” Our Codex usage limit guide covers what to do then. The extension “does not change the model used by Cursor Chat, Agent, Tab, Auto, Cloud Agents, or other Cursor services,” so Cursor’s own agent stays out of usage. And OpenAI calls its Cursor guidance “directional”: “Cursor might change the functionality and availability of these features over time.”

OpenCode, with free or local models

Best for: routine work with no bill and no plan to run out.

OpenCode is an open-source coding agent, and its docs say it “integrates with VS Code, Cursor, or any IDE that supports a terminal.” Its homepage says “OpenCode is 100% free to use. It also comes with a set of free models,” and its providers page adds “You can configure opencode to use local models through Ollama.” Our local model guide covers the setup and the memory it takes.

Price: free, with its free models or a model on your own computer. “There might be additional costs if you connect any other provider.”

Watch for: its free models are free “for a limited time,” and for several of them, “The team is using this time to collect feedback and improve the model.” It’s a different agent, so it starts a new conversation, and what runs well locally depends on how much memory your computer has.

Enjoy

Best for: several labs’ agents in one place, so when one runs out, another carries on.

Enjoy runs Claude Code, Codex, Grok Build, Antigravity, Muse Code, and local models through OpenCode, on Mac, Windows and Linux, on the plans you already have. It doesn’t run Cursor, so it won’t refill your pool; it’s for moving the work to agents with limits of their own. What it adds:

• Another agent, mid-conversation. Hand a job from Claude Code to Codex: “The new agent gets the conversation so far, including replies from other agents, before it works on your message.”
• Limits handled. When an agent runs out, Enjoy switches to another of your accounts with usage left, or waits for the reset and carries on.
• Usage in view. It shows what each signed-in account has left and when it resets, for the agents that report it.

Price: Enjoy is one monthly subscription, and the first seven days are free, with a card to start. It includes your phone and any browser and your other computers, and each teammate you invite needs a subscription of their own. See pricing. The agents are billed separately: Claude Code needs a paid Claude plan and Codex in Enjoy needs ChatGPT Plus or above, while Antigravity works with a free Google account and OpenCode with free or local models.

Watch for: Enjoy isn’t an editor in Cursor’s sense. Its code editor doesn’t suggest code as you type: the agents write the code. It doesn’t run Cursor’s agent or Composer, and it can’t import your Cursor chats, so you start a new conversation. On The X Subscription, Grok Build draws on the same pool that ran out, so it won’t help. Waiting out a limit needs “Your computer must be on, with Enjoy open, when the usage resets,” and scheduled workflows also run only while it’s awake with Enjoy open. Enjoy isn’t open source, and it’s a monthly subscription on top of your AI plans once the free week ends. If you only want Claude or GPT beside Cursor, the two extensions above don’t need Enjoy at all.

When to stay on your plan and wait

If your week resets in a day or two, waiting is free. If you run out once in a while, a bigger plan is money spent for a rare afternoon: pay for that day with credit, and set a cap first. Upgrade when it happens most weeks. And if Tab is what you’d miss, stay in Cursor whichever agent you use: “Tab completion continues using Cursor’s built-in models,” and the extensions above run beside it.

Before you pay for more, check what spent the last lot. Cursor’s usage page says to “Check your usage dashboard for request-level cost details.” Fast mode, third-party subagents and Grok Bot are the first places to look. If you’re on a legacy Pro, Pro+ or Ultra plan and use Claude through its Other Models pool, think before moving to The X Subscription: that allowance goes, and you can’t come back.

How we compared

We read Cursor’s usage, pricing, models, billing, spend limit, API key, rules, legacy plan and X Subscription pages, its security notice and three threads on its forum; SpaceXSI’s usage limits and prepaid balance docs; Anthropic’s pricing, help center and Claude Code docs; OpenAI’s pricing, help, import and Codex docs; and OpenCode’s homepage and docs, all on November 3, 2026, and every one is linked below. We quote each company’s own words for its limits and leave out what it hasn’t documented, such as the size of Cursor’s weekly pool. Plans and limits change often, so we recheck this guide against its sources at least once a month, and the date at the foot of the page says when we last did.