IonBin
All pages
Search
Concept

Rate limits

Rate limits cap API request or token throughput over a period. Limits vary by model, project, organization and account tier. Temporary throttling needs bounded retries; exhausted spend caps or billing problems need separate handling. Consumer-app allowances are separate.

Verified Oct 18, 2026 Review by Apr 16, 2027 3 sources

What it is

A rate limit is a usage ceiling set by an API provider. OpenAI and Google measure limits in requests and tokens per minute and per day. Limits depend on the model and your usage tier.

How it works

  1. Each organization or project has limits per model. OpenAI sets them at the organization and project level, not per user.
  2. OpenAI upgrades tiers as credit purchases reach thresholds. Anthropic considers usage history and account standing; its standard tiers have monthly caps, while Custom arrangements differ.
  3. When you exceed a limit, the API returns an error (HTTP 429) instead of a result. Anthropic uses a “token bucket”, so capacity refills continuously rather than resetting at fixed times.
  4. For temporary rate errors, honor Retry-After when present, then use bounded retries with increasing delays and jitter. Do not retry billing or exhausted monthly caps as if they were temporary congestion. Suitable nonurgent jobs can use batch processing, which has its own limits.

Benefits

  • Knowing your limits helps size queues and handle busy periods.
  • Helps size projects before launch.
  • Explains why a workflow works in testing but fails at scale.

Limits and risks

  • Limits differ by model, vendor and account tier, and change over time. Check your console rather than a published table.
  • Shared organization or project quotas can be exhausted across tools, even with separate keys. Google applies limits per project, not per key.
  • Consumer app limits are separate from API limits. Published ceilings are not a guarantee of available throughput.

Best-fit stacks

Stack 08

Dev, web & IT

Design, code, host and ship sites and internal tools, with an AI coding agent doing most of the typing.

Agent compatibility

Not tool-specific. See Agent-ready for connector support by tool.

Levels come from each tool's recorded, sourced compatibility. How the levels work.

Sources

  1. Rate limits · OpenAIofficial accessed Oct 18, 2026
  2. Rate limits · Anthropicofficial accessed Oct 18, 2026
  3. Rate limits · Googleofficial accessed Oct 18, 2026

Practitioner views are restated in our words and labeled as views, not measured results. Product facts come from official sources dated above.

Suggest an edit

IonBin's wiki isn't user-edited. Corrections go to our Directory Curator, who checks them against primary sources before anything changes. Spot something wrong or out of date on “Rate limits”? Email us with a link to a primary source.

Suggest an edit by email or write to contact@ionbin.com