# Grok Bot Usage Limits Explained: The Weekly Pool, the November Throttling, and Your Options > How Grok Bot usage limits work: the weekly pool shared with Grok and Cursor, Premium, Super and Ultra plans, what changed in November 2026, and your options when a bot runs out. Verified November 8, 2026. Published November 8, 2026 · by the AQ team · Canonical: https://aq.dev/guides/grok-bot-usage-limits/ Grok Bot has no separate quota of its own. On The X Subscription, each paid seat gets one weekly usage pool that Grok Bot shares with the Grok apps and Cursor, and the pool's size depends on your plan: Premium ($30 per month), Super ($100) or Ultra ($200), per xAI's documentation. When the pool runs out, further Grok Bot usage is paid from prepaid credit, or the bot waits for the weekly refresh. In early November 2026, many paying users reported reaching their weekly limit days early, and Elon Musk said on November 1 that the cost of serving Grok Bot had been unsustainably high relative to revenue. This guide uses xAI's own pages as the source for plans and mechanics, verified on November 8, 2026, and marks clearly where news reports and user accounts go further than xAI has published. ## How does the Grok Bot weekly pool work? xAI's [usage limits page](https://docs.x.ai/x-subscription/usage-limits) says each paid seat on The X Subscription "receives a weekly pool for eligible first-party usage only across supported Grok and Cursor surfaces," and that the pool is shared "rather than split into separate product quotas." The [Grok Bot FAQ](https://docs.x.ai/grok-bot/faq) confirms that "Grok Bot draws from the same weekly usage pool as Cursor and the Grok apps." - **One meter per seat.** A long Grok Bot task, a Cursor session and a Grok chat all draw from the same pool. Heavy use in one leaves less for the others that week. - **Different rates per product.** xAI says "usage rates vary by product, so different activities can draw different amounts from the pool." It does not publish the rates. - **Weekly refresh.** Cursor's [Grok Bot plans page](https://cursor.com/help/grok-bot/plans) says "the pool refreshes weekly." Your live usage view shows the next reset date. xAI does not publish a fixed reset day for every account. - **After the pool.** Further first-party usage becomes Additional Usage, paid from prepaid credit bought with top-ups. Auto Top-up and a monthly spending cap for Additional Usage are set under Additional usage credits in your billing settings. - **Third-party models in Cursor** are paid Additional Usage and sit outside the weekly pool. ### How to check how much Grok Bot usage you have left Open **grok.com/dashboard** and select **Usage**. xAI says the live usage view shows how much of the pool you have used and the next reset date. The billing view shows prepaid credit, Auto Top-up and your spending cap. If you are still on a legacy Cursor plan, Grok Bot usage is metered on your Cursor account instead. ## Grok Bot plans and prices xAI's [X Subscription pricing page](https://docs.x.ai/x-subscription/pricing) lists these tiers. Grok Bot is included on Premium, Super and Ultra; Plus and Free do not include it. Team seats cost the same per seat and include the same usage as the matching individual tier. | Plan | Monthly | Annual | Usage vs Plus | | --- | --- | --- | --- | | Free | $0 | $0 | Limited | | Plus | $8 | $80 | Base | | Premium | $30 | $300 | 4x | | Super | $100 | $1,000 | 15x | | Ultra | $200 | $2,000 | 50x | Prices are USD as listed by xAI on November 8, 2026; check [x.ai/pricing](https://x.ai/pricing) for your region and the current terms. xAI publishes multiples, not token or task counts, so "4x" tells you how plans compare with each other and nothing about how many Grok Bot tasks a week you will get. Secondary sources express the same ladder differently. Several news reports, citing xAI's Grok Bot page, describe Super as 3.5 times Premium usage and Ultra as 12 times Premium. Those figures are close to the documented 15x and 50x of Plus divided by Premium's 4x, but they are rounded and come from a different page, so treat the docs table above as the reference. An earlier xAI announcement from August 26, 2026, before The X Subscription launched in October, described Grok Bot usage as separate from Grok and Cursor plans. The current docs say the pool is shared, and that is what applies today. ### Do you need a subscription for Grok Bot? Yes. Grok Bot comes with Premium, Super and Ultra on The X Subscription, and with the legacy Cursor Pro, Pro+, Ultra and Teams plans. Free and Plus do not include it. Cursor's help page also mentions a free trial on legacy Cursor plans, given as a usage credit with a 7-day window. ## What changed in November 2026? xAI has not published old and new limits, so what follows is a timeline of what was said and reported, not a measured change. - **Late October to November 1.** Paying users reported weekly meters jumping to 100% mid-week, some saying they hit the limit with days left before the reset. These are user reports; their numbers are not independently verified. - **November 1.** Replying on X to a user asking about a usage reset, Elon Musk wrote that "the cost of serving Grok Bot was unsustainably high relative to revenue," as [reported by TeslaNorth](https://teslanorth.com/2026/11/01/elon-musk-grok-bot-cost-usage-limits/) and [Forbes](https://www.forbes.com/sites/siladityaray/2026/11/02/musk-says-grok-bots-costs-were-unsustainably-high-weeks-after-touting-rapid-growth/). He did not say what changed or whether prices would move. - **November 8.** Musk posted that "Grok Bot fixes for improved token efficiency and smarter behavior are underway" and said he would look into one CTO's report that his allowance lasted 14 hours under moderate load ([TeslaNorth](https://teslanorth.com/2026/11/08/grok-bot-usage-limits-musk/)). No timeline was given, and the post does not promise a larger weekly allowance. Some coverage has linked part of the meter jumps to a billing problem on the Cursor side. Neither xAI nor Cursor has confirmed that publicly, so it remains unexplained. If your meter moved while you were not using Grok Bot, remember the pool is shared: Cursor sessions and Grok chats on the same seat count too. If the numbers still look wrong, raise it with xAI support before buying more usage. ## Why bots burn through a weekly pool faster than chat A chat turn is one request and one answer. A bot does many steps per task, and each step carries the growing context of everything before it. A few patterns account for most of the difference: - **Unattended routines.** A routine that runs every hour, or fires on every new email, keeps spending while you are away. Twenty small runs a day add up across a week. - **Browser and computer use.** xAI describes Grok Bot as having its own always-on cloud computer with browser and terminal access. Every screenshot and page read goes back into the model's context, which costs far more than plain text. - **Long tasks.** A multi-hour job re-reads its accumulated history on every step, so the cost of each step grows as the task goes on. - **Retries.** A task that fails and starts over pays for the same work twice. ## What can you do when Grok Bot hits its limit? - **Check the meter first.** Look at Usage on grok.com/dashboard for the remaining pool and reset date, and check whether Cursor or Grok chat on the same seat used it. - **Give each run less to do.** Narrow the instruction, cut the pages a bot must read, and ask for a short report instead of a full walkthrough. - **Pause or space out routines.** Move hourly jobs to daily, and stop the ones whose output nobody reads. - **Wait for the weekly refresh** if the remaining work can wait. With Auto Top-up off, nothing is charged past the pool. - **Add prepaid credit with a cap.** Top up Additional usage credits and set a monthly limit so a busy week cannot run past your budget. - **Move to a larger plan.** Super and Ultra include larger weekly pools than Premium. Compare the multiples above with how far into the week you get today. - **Run standing jobs where the model bill is one you already pay.** If your team already pays for Claude or ChatGPT plans, routines that run on those subscriptions use limits you are already paying for. For the broader billing trade-off, see [subscription vs API credits](https://aq.dev/guides/subscription-vs-api-credits-for-coding-agents/). For how other vendors meter agents, see [Claude Code usage limits](https://aq.dev/guides/claude-code-usage-limits-explained/) and [Codex usage limits](https://aq.dev/guides/codex-usage-limits-explained/). ## Where AQ fits AQ bots run routines on schedules and events with a record of every run and a monthly cap, and they can run on your own connected Claude or OpenAI subscription, where AQ never marks up the usage. AQ is the multiplayer coding harness where teams build and run bots with every AI coding agent, like Claude Code and Codex. A bot in AQ is a named agent with one ongoing chat per person. Its routines each have an instruction and one trigger: a schedule, a monthly date, every N minutes, a Linear label, Gmail, Google Calendar, GitHub events, a webhook, or a manual run. Each run has a record of why it woke, what it did, what it produced and what it cost, and a routine can post to the chat only when something happened. On money: the bot's monthly cap covers the AQ credits its routines spend, and an optional per-run ceiling stops a single run. A routine runs as the person who created it, on their connected subscriptions or on AQ models at the per-token price shown before you pick one. AQ credits are priced, and a connected Claude or OpenAI subscription keeps its vendor's own limits, so a bot in AQ still has limits; they are ones you can see and set. When the answer is code, a bot starts a real work item with a branch and a pull request. Read [how AQ bots work](https://aq.dev/docs/bots/), see [AQ Bots](https://aq.dev/product/bots/), or [connect Grok Bot to AQ](https://aq.dev/guides/connect-grok-bot-to-aq/) if you want to keep Grok Bot and hand coding work to AQ. The Free plan includes a personal sandbox machine; the [AQ Pro plan](https://aq.dev/pricing/) is $50 per user per month during early access, and each Pro seat includes AQ credits every month. ## Frequently asked questions ### What are Grok Bot's usage limits? Grok Bot draws from one weekly usage pool per paid seat, shared with the Grok apps and Cursor. The pool's size depends on the plan: xAI lists Premium at 4x Plus usage, Super at 15x and Ultra at 50x. xAI does not publish task or token counts. Past the pool, usage is paid from prepaid credit. ### How many chats can you have with Grok? xAI does not publish a fixed number of chats or Grok Bot tasks. Usage comes out of a weekly pool whose size depends on your plan, and different activities draw different amounts from it. The Usage view on grok.com/dashboard shows how much you have used and when the pool refreshes. ### Do you need a subscription for Grok Bot? Yes. Grok Bot is included on Premium, Super and Ultra on The X Subscription, and on the legacy Cursor Pro, Pro+, Ultra and Teams plans. Free and Plus do not include it. ### How much does Grok Bot cost? Grok Bot comes with The X Subscription from Premium up: Premium is $30 per month ($300 a year), Super $100 ($1,000) and Ultra $200 ($2,000), per xAI's pricing docs on November 8, 2026. Team seats cost the same per seat. Usage past the weekly pool is paid from prepaid credit. ### How do I get more Grok Bot usage? Wait for the weekly refresh, add prepaid Additional usage credits (with Auto Top-up and a monthly cap if you want), or move to Super or Ultra for a larger weekly pool. Giving each run less to do and pausing busy routines also stretches the pool you have. ### Is the Grok weekly limit shared with Cursor? Yes, on The X Subscription. xAI's FAQ says Grok Bot draws from the same weekly pool as Cursor and the Grok apps. On a legacy Cursor plan, Grok Bot usage is metered on the Cursor account instead. ### Did xAI cut Grok Bot limits in November 2026? xAI has not published old and new limits. Users reported reaching weekly limits early, Elon Musk wrote on November 1 that the cost of serving Grok Bot had been unsustainably high relative to revenue, and on November 8 he said fixes for token efficiency were underway, without a timeline. ### Can AQ bots run on my existing Claude or OpenAI subscription? Yes. An AQ routine runs as the person who created it, on their connected subscriptions or on AQ models at published per-token prices. AQ never marks up usage on a connected subscription, and that subscription keeps its vendor's own limits. A bot's monthly cap covers the AQ credits its routines spend. --- About AQ: AQ is the multiplayer coding harness where teams build and run bots with every AI coding agent, like Claude Code and Codex: shared live terminals, review and merge, app previews and standing jobs, in your own cloud. Start at [aq.dev](https://aq.dev/).