Weekly pool
Shared SuperGrok pool
Fact
xAI · Subscription and API limits
Grok Build Usage Limits
Grok Build draws from a shared weekly pool on paid Grok plans. A fixed Build prompt or token allowance is not publicly disclosed; API throughput limits describe a separate system.
- Last verifiedSep 7, 2026, 6:30 PM UTC
- Official sources on this page5
- Limit systemsWeekly pool and API rates
- Numeric Build quotaUnknown
Reset
Settings → Usage
Fact
Numeric Build quota
Not published
Unknown
API limits
Separate system
Fact
Direct answer
What is the weekly Grok Build limit, and when does it reset?
Paid Grok Build uses the shared SuperGrok weekly pool, not a published prompt or token count. Open Settings → Usage for your reset time; Extra Usage Credits can continue after the pool is empty.
Verified fact breakdown
- Current plans
- Fact Paid Grok / SuperGrok subscription usage, or a separate xAI API account. Subscription prices are plan-dependent and not stored as one Build price.
- Primary quota
- Fact One shared weekly usage pool across Grok products, including Build.
- Weekly quantity
- Unknown xAI confirms a weekly pool exists, but does not publish a fixed numeric Grok Build prompt or token quota.
- Reset
- Fact The pool resets on the schedule shown in Settings → Usage. TraceCheck does not invent a UTC reset.
- After the limit
- Fact Paid features pause until reset. Extra Usage Credits, a plan upgrade, or Auto Top Up can continue paid use. Free-tier Chat and Voice remain available.
- Last verified
- Sep 7, 2026, 6:30 PM UTC
Reset windows
When does SuperGrok reset?
A timeline of official windows, not remaining quota. Unknown stays unknown.
Weekly SuperGrok pool
Fact
The pool renews each week. Open Settings → Usage for the reset date and time associated with your account, Extra Usage Credits balance, and the product breakdown.
Universal reset clock
Unknown
These sources do not establish one universal reset day or hour for every subscriber. TraceCheck does not invent a UTC reset.
API RPS / TPM
Fact
API rate limits apply continuously. They are not a weekly SuperGrok reset and do not consume that pool.
See the official usage and limits FAQ for the account usage display.
Need to continue now? Check the reset timestamp and Extra Usage Credits balance in Settings → Usage. Waiting for the displayed reset avoids buying continuation; paid continuation has its own charge. If your entitlement came through Cursor, first check whether you are using Grok Bot or Grok Build.
Published structure
Published limits at a glance
One table for subscription and API disclosures. Unknown is not a remaining-quota bar.
| System | Measure | Officially documented state |
|---|---|---|
| Subscription / SuperGrok | Shared weekly pool | Yes — shared across Grok products, including Build |
| Fixed Grok Build numeric quota | Unknown — not publicly disclosed | |
| Reset model | Weekly pool; the reset schedule is shown in Settings → Usage | |
| Free-plan Grok Build numeric quota | Unknown — not publicly disclosed | |
| API / grok-build-0.1 | Tier 0 request rate | 37 requests/second (RPS) |
| Tier 0 token rate | 10,000,000 tokens/minute (10M TPM) | |
| Token pricing | $1 input / $0.20 cached input / $2 output per 1M tokens | |
| Context window | 256,000 tokens |
How each published cap is scoped
Subscription
Shared weekly SuperGrok pool
Fact
Paid Grok Build draws from one weekly pool shared across Grok products, including Build. Different products consume the pool at different compute costs.
Subscription
Fixed Grok Build numeric quota
Unknown — not publicly disclosed
Unknown
xAI confirms a weekly pool exists, but does not publish a fixed numeric Grok Build prompt or token quota.
Subscription
Free-plan Grok Build numeric quota
Unknown — not publicly disclosed
Unknown
Free-tier Chat and Voice remain on their own schedule after the paid weekly pool is empty. A free-plan Build number is not published.
API / grok-build-0.1
Tier 0 request rate
37 requests/second (RPS)
Fact
This is a continuous API rate limit, not the SuperGrok weekly pool.
API / grok-build-0.1
Tier 0 token rate
10,000,000 tokens/minute (10M TPM)
Fact
Exceeding RPS or TPM returns HTTP 429. Tiers never downgrade.
API / grok-build-0.1
Context window
256,000 tokens
Fact
Maximum model context for one API request, not an account quota and not a remaining-usage meter.
After the limit
What happens when I hit the weekly limit?
Canonical branches only. No invented remaining-quota path.
If the weekly pool is empty
Paid features pause until the weekly pool resets. Extra Usage Credits, a plan upgrade, or Auto Top Up can continue paid use. Free-tier Chat and Voice remain available on their own schedule.
Fact
Wait for reset
Paid features pause until the weekly pool resets. Check the timestamp in Settings → Usage instead of assuming a global clock.
Fact
Use Extra Usage Credits
Extra Usage Credits can continue paid features after the weekly included pool is exhausted. Credits can currently be purchased on the web, are used after included weekly usage, and expire one year after purchase unless otherwise stated. Auto Top Up is optional.
Fact
Stay on free Chat and Voice
Free-tier Chat and Voice remain available on their own schedule after the paid weekly pool is empty.
Derived: Extra Usage Credits are a continuation path, not a published numeric Build quota and not an API rate limit. API HTTP 429 is a separate blocked path and does not consume or reset the SuperGrok weekly pool.
Plan comparison
Subscription pool vs API rate limits
Units stay as published. RPS, TPM, and the weekly pool are not merged into one score.
| Plan | System | Quota | Published quantity | Reset | After the limit |
|---|---|---|---|---|---|
| Grok / SuperGrok subscriptionShared weekly pool | Shared weekly pool | Fact One weekly usage pool across Grok products, including Build | Unknown Unknown — not publicly disclosed | Weekly; schedule shown in Settings → Usage | Paid features pause until reset. Extra Usage Credits, a plan upgrade, or Auto Top Up can continue paid use. |
| xAI APIgrok-build-0.1 rate limits | grok-build-0.1 rate limits | Fact Per-model RPS and TPM, separate from SuperGrok | Fact Tier 0: 37 RPS and 10,000,000 TPM | Continuous API rate limit | Exceeding per-model RPS or TPM returns HTTP 429. This does not consume or reset the SuperGrok weekly pool. |
Freshness
Recent verified changes
Public verified records only. Pending candidates are not shown.
No verified Grok Build change records are in the public feed yet.
Metering map, usage decoder, and source conflicts
Metering map
What Grok Build actually meters
Each plan below is a pool graph: consumers drain a pool, a reset clock refills it, and exhaustion says what happens at empty. Numeric capacity is shown only when an official source publishes it for that exact pool.
Grok / SuperGrok subscription
UnknownShared weekly SuperGrok pool · shared · usage value
xAI confirms a weekly pool and a percentage display with a per-product breakdown. A fixed numeric Grok Build prompt or token quota is not published.
- Consumer · surface Grok Build — Fact Build draws from the shared weekly SuperGrok pool. A long coding task uses more of the pool than a short chat.
- Consumer · surface Grok Chat — Fact Chat draws from the same weekly pool at a different compute cost.
- Consumer · surface Grok Imagine — Fact Imagine draws from the same weekly pool. High-quality video uses more compute than a short chat.
- Consumer · surface Grok Voice — Fact Voice draws from the same weekly pool.
- Consumer · surface API line in Settings → Usage — Unknown Settings → Usage lists API in the weekly product breakdown. TraceCheck does not treat that display as the developer API RPS/TPM table.
- Reset · weekly Fact The weekly pool resets on the account schedule shown in Settings → Usage. TraceCheck does not invent one UTC reset for every subscriber. Visible in Settings → Usage.
- At limit · wait for reset, credits, upgrade Fact Paid features pause until the weekly pool resets. Extra Usage Credits, a plan upgrade, or Auto Top Up can continue paid use. Free-tier Chat and Voice remain available on their own schedule.
Source: Grok usage and limits FAQ · verified Sep 7, 2026, 6:30 PM UTC
xAI API
Factgrok-build-0.1 API RPS · requests
Published capacity: 37.
- Consumer · request type grok-build-0.1 API requests — Fact Tier 0 grok-build-0.1 is limited to 37 requests per second. Exceeding the cap returns HTTP 429.
- Reset · unknown Fact API RPS is a continuous rate limit, not a weekly SuperGrok reset.
- At limit · blocked Fact Exceeding per-model RPS returns HTTP 429. This does not consume or reset the SuperGrok weekly pool.
Source: xAI API rate limits · verified Sep 7, 2026, 6:30 PM UTC
Factgrok-build-0.1 API TPM · tokens
Published capacity: 10000000.
- Consumer · request type grok-build-0.1 API tokens — Fact Tier 0 grok-build-0.1 is limited to 10,000,000 tokens per minute. Exceeding the cap returns HTTP 429.
- Reset · unknown Fact API TPM is a continuous rate limit, not a weekly SuperGrok reset.
- At limit · blocked Fact Exceeding per-model TPM returns HTTP 429. This does not consume or reset the SuperGrok weekly pool.
Source: xAI API rate limits · verified Sep 7, 2026, 6:30 PM UTC
Usage decoder
Decode your usage
One compact answer path from the verified metering graph. TraceCheck will not guess which clock you hit.
Do Grok Build and Grok Chat share usage?
DocumentedShared pools: Shared weekly SuperGrok pool. Only documented consumers on those pools draw the same allowance. A surface not listed as a consumer is not assumed to share it.
Mechanics
- Shared weekly SuperGrok pool consumers: Grok Build, Grok Chat, Grok Imagine, Grok Voice, API line in Settings → Usage.
- Grok Build — Coding product listed in the SuperGrok weekly usage breakdown.
- Grok Chat — Consumer chat on grok.com and the Grok apps.
- Grok Imagine
Unknown / conflict
- Unknown The verified sources do not establish that xAI API shares this pool: Developer API with per-model RPS/TPM. Not the SuperGrok weekly quota.
- Unknown The verified sources do not establish that Grok Bot shares this pool: Separate durable-bot product. Not grok.com Grok and not assumed to drain the Build weekly pool.
Source: Grok usage and limits FAQ · Grok plans and usage overview
What counts against Grok Build usage?
PartialShared weekly SuperGrok pool is drained by Grok Build, Grok Chat, Grok Imagine, Grok Voice, API line in Settings → Usage. grok-build-0.1 API RPS is drained by grok-build-0.1 API requests. grok-build-0.1 API TPM is drained by grok-build-0.1 API tokens.
Mechanics
- Grok / SuperGrok subscription · Shared weekly SuperGrok pool: Build draws from the shared weekly SuperGrok pool. A long coding task uses more of the pool than a short chat.
- Grok / SuperGrok subscription · Shared weekly SuperGrok pool: Chat draws from the same weekly pool at a different compute cost.
- Grok / SuperGrok subscription · Shared weekly SuperGrok pool: Imagine draws from the same weekly pool. High-quality video uses more compute than a short chat.
- Grok / SuperGrok subscription · Shared weekly SuperGrok pool: Voice draws from the same weekly pool.
Unknown / conflict
- Unknown xAI confirms a weekly pool and a percentage display with a per-product breakdown. A fixed numeric Grok Build prompt or token quota is not published.
Source: Grok usage and limits FAQ · Grok plans and usage overview · xAI API rate limits · Grok Build 0.1 model documentation
Is the API quota the same as the weekly subscription pool?
DocumentedNo. The SuperGrok weekly subscription allowance and the developer API's per-model request/token rate limits are separate systems. An API HTTP 429 does not identify exhaustion of the weekly subscription pool.
Mechanics
- Shared weekly SuperGrok pool uses usage value.
- grok-build-0.1 API RPS uses requests.
- grok-build-0.1 API TPM uses tokens.
Source: Grok usage and limits FAQ · Grok plans and usage overview · xAI API rate limits · Grok Build 0.1 model documentation
Is Grok Build the same as Grok Bot?
DocumentedNo. Grok Build is the coding product in the SuperGrok weekly usage breakdown. Grok Bot is a separate durable-bot product, not Grok on grok.com or the Grok mobile apps. Do not treat their allowances as interchangeable.
Source: Grok usage and limits FAQ · Grok Bot FAQ
Does Cursor's Grok access use the same SuperGrok pool?
PartialCursor access is not evidence of a SuperGrok Build weekly-pool entitlement. Cursor documents a separate Grok Bot usage grant; that does not establish the pool used by every Grok model or provider route in Cursor.
Next action: Check whether you are using Grok Bot, a Cursor-included model, or your own API key before comparing usage meters.
Unknown / conflict
- Unknown — undisclosed The exact pool for an unspecified Cursor model/provider route is not established here. Official xAI and Cursor sources also disagree on the eligible Grok Bot plan list; see the conflict below.
Source: Grok usage and limits FAQ · Grok Bot FAQ · Cursor Grok Bot plans and billing
Diagnose this in the calculator · Grok Bot vs Grok Build: Access, Usage Pools, and Limits
Official source conflict
Official sources disagree
Unknown — conflicting Which SuperGrok and Cursor plans currently include Grok Bot
- Source A: Grok Bot FAQ — The Grok Bot FAQ lists eligible plans as SuperGrok Plus, SuperGrok Heavy, Cursor Pro+, Cursor Ultra, and Cursor Teams Standard and Premium. If both Cursor and SuperGrok subscriptions exist, Grok Bot uses whichever has more usage.
- Source B: Cursor Grok Bot plans and billing — Cursor help says Grok Bot is included on every paid individual Cursor plan and Cursor Teams, and that linking individual SuperGrok, SuperGrok Plus, SuperGrok Heavy, or X Premium+ is a usage grant on the Cursor account. SuperGrok Lite, SuperGrok Team, and SuperGrok Enterprise are not included.
- Applies to: Users deciding whether a current SuperGrok or Cursor plan includes Grok Bot
- Decision impact: Do not treat one eligible-plan list as settled Fact. Build vs Bot pool separation does not depend on resolving this list. Cursor-included Bot still does not imply SuperGrok Build weekly-pool entitlement.
- TraceCheck status: Unresolved — Derived outputs must not treat one side as settled Fact.
- Current interpretation: TraceCheck records that official Bot plan lists currently disagree. The August 26 xAI news post lists SuperGrok, Cursor Pro, and Cursor Teams, which matches Cursor help more closely than the Bot FAQ list and remains a third official description.
- Last checked Sep 8, 2026, 8:00 PM UTC
Which Grok surface uses which pool?
Which Grok surface uses which pool?
| Surface | Included by | Consumes this pool? | Notes |
|---|---|---|---|
| Grok Build | Paid Grok / SuperGrok | Fact Shared weekly SuperGrok pool | Listed in the Settings → Usage product breakdown. |
| Grok Chat, Imagine, Voice | Paid Grok / SuperGrok | Fact Same shared weekly pool | Different products cost different amounts of compute. |
| Free-tier Chat and Voice | Free Grok | Fact Separate free-tier limits | Remain available on their own schedule after the paid weekly pool is empty. |
xAI API grok-build-0.1 | API account | Fact API RPS/TPM, not the weekly SuperGrok quota | Tier 0 is 37 RPS and 10M TPM. HTTP 429 on exceed. |
| Grok Bot | Eligible SuperGrok or Cursor plans — official lists currently conflict | Fact Separate weekly Bot usage, not the SuperGrok Build pool | See Grok Bot vs Grok Build. Cursor-included Bot is not a SuperGrok Build entitlement. Numeric Bot capacity is Unknown. |
| Cursor-included Grok models | Cursor plans | Not assumed | Cursor Grok access is not copied into the SuperGrok Build weekly pool in this record. |
Background: weekly pool, tokens, API, and mixed numbers
Grok weekly limits and the shared Build quota
Fact Paid users receive one weekly allowance that can be spent across Grok products. Build appears in xAI's usage breakdown, and higher paid tiers provide larger weekly allowances. Different products consume the pool at different compute costs.
xAI's verified documentation does not provide a stable number of Grok Build tasks, prompts, tokens, or compute units included in that pool. The fixed subscription cap therefore remains unknown / not publicly disclosed.
Are token limits or CLI percentages a Build quota?
A fixed subscription token allowance is Unknown. An observed CLI percentage cannot establish a universal number of remaining prompts or tokens. The API model's context window describes request capacity; its token rate describes throughput.
Derived Identify whether work uses subscription access or an API account before interpreting a limit. A SuperGrok subscription quota cannot be read from the API rate table.
xAI API usage is separate
Fact The API uses metered token pricing and per-model rate limits. For grok-build-0.1, xAI lists a 256,000-token context window, Tier 0 limits of 37 RPS and 10M TPM, and prices of $1.00 per 1M input tokens, $0.20 per 1M cached input tokens, and $2.00 per 1M output tokens.
These API limits do not state how much Grok Build usage is included with a Grok or SuperGrok subscription. Exceeding API RPS or TPM returns HTTP 429. Exceeding per-model RPS or TPM returns HTTP 429. This does not consume or reset the SuperGrok weekly pool.
Why you may see different numbers online
- User-measured percentages and CLI displays describe observed account usage; they are not canonical fixed limits.
- A 256,000-token context window is the maximum model context for one API request, not an account quota.
- API RPS and TPM rate limits control request throughput; they are not subscription usage limits.
Official sources conflict
Official sources conflict Official xAI Grok Bot FAQ, Cursor Grok Bot plans help, and the August 26 xAI news post currently describe different eligible SuperGrok and Cursor plan sets for Grok Bot.
Verified Sep 8, 2026, 8:00 PM UTC. Affected pages: /grok-build/usage-limits/, /grok-bot-vs-grok-build/
Official evidence
Official sources
Canonical vendor pages only. Observed CLI percentages are not sources.
Related tasks