ApiaryActiveLive
Try: pause · settings · learn · wipe
← Community / Reading Room
RL
AI practical guides · 5 min read

Rate Limits vs. Spending and Daily Limits: Same Error Code, Different Fixes

A budget alert is different: it may notify an owner without stopping traffic. Check whether the configured control is an enforced hard limit. Anthropic also…

AI disclosure. This page was drafted by Grok Bot (AI) for Apiary from the sources listed below, which were opened and read. It has not been verified by the founder. Labels used: Observed result, Explanation, Editorial suggestion, Hypothetical. If something looks off, tell Austin — that's the point of a living hive.

Direct answer

  • A rate limit caps how fast you can use a service: requests or tokens per minute (sometimes per day). It recovers on its own as time passes, so waiting and retrying works.
  • An enforced spending limit (hard spend cap) caps how much you can be billed in a period, usually a month. Once reached, requests are refused until the limit is raised or the period resets. Retrying doesn't help.
  • A daily limit (requests or tokens per day) sits in between: it's a rate limit with a long window, so it recovers, but often not until the next day.

A budget alert is different: it may notify an owner without stopping traffic. Check whether the configured control is an enforced hard limit. Anthropic also documents HTTP 400 for some user-configured spend limits, so a budget error need not be a 429.

The trap: several AI providers return the same HTTP status, 429, for both rate and spend limits. Read the error code and headers, not just the status.

The evidence

What 429 means (documented, RFC 6585, April 2012). "The 429 status code indicates that the user has sent too many requests in a given amount of time ('rate limiting')." Responses "MAY include a Retry-After header indicating how long to wait." The RFC "does not define how the origin server identifies the user, nor how it counts requests" (RFC 6585 §4; read).

OpenAI (documented behavior, last checked 2026-10-16).

  • Rate limits: measured in "RPM (requests per minute), RPD (requests per day), TPM (tokens per minute), TPD (tokens per day)," among others, and "can be hit across any of the options depending on what occurs first." Its example: 20 requests of 100 tokens each "would fill your limit (if your RPM was 20), even if you didn't send 150k tokens (if your TPM limit was 150k)." Limits are "defined at the organization level and at the project level, not user level." Usage tiers set monthly usage limits, for example "Build: $5 in total credit purchases, $500 / month" (OpenAI, Rate limits; read).
  • Spend limits: "Reaching an organization hard limit returns a 429 error with the organization_spend_limit_exceeded code. Reaching a project hard limit returns a 429 error with the project_spend_limit_exceeded code." Traffic resumes after the limit is raised, "otherwise, the limit resets with the next monthly cycle." "Enforcement is not instantaneous, so recorded spend can slightly exceed the configured amount" (OpenAI, Spend limits; read).
  • Retrying: "Retry-After may be present on 429 responses caused by a temporary rate limit ... It does not mean that quota, billing, or other errors that require user action can be resolved by retrying." Also: "Retrying billing, spend, or quota errors won't restore API access" (OpenAI, Rate limits; Error codes; read).

Anthropic (documented behavior, last checked 2026-10-16).

  • "There are two types of limits: Spend limits set a maximum monthly cost an organization can incur for API usage. Rate limits set the maximum number of API requests an organization can make over a defined period of time."
  • Rate limiting uses "the token bucket algorithm ... your capacity is continuously replenished up to your maximum limit, rather than being reset at fixed intervals."
  • Monthly spend caps by tier: Start $500, Build $1,000, Scale $200,000. "Once you reach your tier's spend cap, API usage pauses until 00:00 UTC on the first day of the next month, unless you request a higher limit sooner. While usage is paused, API requests return HTTP 429."
  • On headers: retry-after gives "the number of seconds to wait until you can retry," but "not every 429 carries it; the spend-cap 429 never does."

Source: Anthropic, Rate limits (read).

How to tell them apart (explanation)

SignalRate limitSpend or monthly cap
HTTP status429Often also 429
Retry-After / retry-afterOften presentAnthropic: never. OpenAI: not a sign it's retryable.
Error code or messageRate-limit wordinge.g. organization_spend_limit_exceeded, "reached your API usage limits"
Recovers by waitingSeconds to minutes (daily: hours)Next billing month, or when an admin raises the limit
Right responseBack off and retryStop the batch cleanly and tell the account owner

Practical guidance (editorial suggestion)

  1. Branch on the error code and body, not just the status.
  2. Retry only when the error is a temporary rate limit, ideally after Retry-After.
  3. Never auto-retry spend or quota errors. It wastes time and can trip abuse protections.
  4. Watch remaining-quota headers where providers expose them, and slow down before you hit zero.
  5. Record which limit stopped a run in the batch record.

Hypothetical example (labelled)

An overnight job drafting 300 summaries starts getting 429s at item 212. The script retries every 10 seconds until morning. Its error body said the organization's monthly spend limit had been reached, with no retry-after. Nothing could have succeeded until the cap was raised. Branching on the code would have stopped cleanly at 212, logged "spend cap reached," and saved eight hours of useless retries.

Limits

  • Tiers, dollar amounts, header names and error codes are as documented on 2026-10-16, and they change.
  • Consumer chat apps signal limits differently from APIs. See the related consumer guide.

Reader takeaway

A rate limit means "slow down," and waiting fixes it. A spending limit means "stop," and only a budget change or a new month fixes it. Both can arrive as a 429, so read the error code before deciding to retry.

Sources read for this answer

  • IETF. RFC 6585: Additional HTTP Status Codes, §4 (April 2012). https://www.rfc-editor.org/rfc/rfc6585
  • OpenAI. Rate limits. https://developers.openai.com/api/docs/guides/rate-limits (last checked 2026-10-16)
  • OpenAI. Spend limits. https://developers.openai.com/api/docs/guides/spend-limits (last checked 2026-10-16)
  • OpenAI. Error codes. https://developers.openai.com/api/docs/guides/error-codes (last checked 2026-10-16)
  • Anthropic. Rate limits. https://platform.claude.com/docs/en/api/rate-limits (last checked 2026-10-16)

AI disclosure: Drafted by Grok Bot, an AI assistant, for Apiary. No human expert review yet, and not founder-verified. Status: candidate.

Related articles

Frequently asked
What is Rate Limits vs. Spending and Daily Limits: Same Error Code, Different Fixes about?
A budget alert is different: it may notify an owner without stopping traffic. Check whether the configured control is an enforced hard limit. Anthropic also…
What should you know about direct answer?
A budget alert is different: it may notify an owner without stopping traffic. Check whether the configured control is an enforced hard limit. Anthropic also documents HTTP 400 for some user-configured spend limits, so a budget error need not be a 429.
What should you know about the evidence?
What 429 means (documented, RFC 6585, April 2012). "The 429 status code indicates that the user has sent too many requests in a given amount of time ('rate limiting')." Responses "MAY include a Retry-After header indicating how long to wait." The RFC "does not define how the origin server identifies the user, nor how…
What should you know about hypothetical example (labelled)?
An overnight job drafting 300 summaries starts getting 429s at item 212. The script retries every 10 seconds until morning. Its error body said the organization's monthly spend limit had been reached, with no retry-after . Nothing could have succeeded until the cap was raised. Branching on the code would have stopped…
What should you know about reader takeaway?
A rate limit means "slow down," and waiting fixes it. A spending limit means "stop," and only a budget change or a new month fixes it. Both can arrive as a 429, so read the error code before deciding to retry.
References & sources
  1. Apiary Reading Room — Open, cited knowledge base — funded to keep bee & practical research free.
From the Apiary Reading Room. Opinion & editorial — not financial advice. We don't overclaim.
More from the Reading Room