API Rate Limit Calculator

What This Calculator Measures

AI providers cap how much you can send them per minute — usually through two separate limits: RPM (requests per minute) and TPM (tokens per minute). This calculator figures out your real maximum throughput given both limits, and tells you which one is actually holding you back.

Why Two Limits Instead of One

RPM caps how many separate API calls you can fire, regardless of size. TPM caps the total token volume across those calls, regardless of how many requests it's split across. A workload with many tiny requests can hit the RPM ceiling while barely touching its token budget; a workload with few huge requests can do the opposite and blow through TPM while sending very few requests. You need to check both.

The Math

max requests/min by RPM = your RPM limit
max requests/min by TPM = floor(your TPM limit ÷ average tokens per request)
effective max/min        = the smaller of the two
binding constraint        = whichever ceiling is lower

Why We Don't Hardcode Provider Tier Tables

Rate limits are tied to your account's usage tier (often based on cumulative spend), and providers adjust them over time. A table baked into this calculator would go stale within months — instead, pull your actual current limits from your provider's dashboard (for example, OpenAI's organization limits page) and plug them in here for a result you can trust.

Common Mistakes

• Sizing infrastructure off request count alone. If your binding constraint is TPM, adding more parallel request workers won't help — you'll just get rate-limited faster.
• Forgetting rate limits are per-organization, not per-API-key. Multiple integrations or team members sharing one account share the same pool of RPM/TPM.

Frequently Asked Questions