Pricing & Cost
Requests Per Minute (RPM)
Quick Answer
A rate limit measuring maximum requests allowed per minute.
RPM is a request frequency limit. RPM limits prevent single users overwhelming infrastructure. Typical limits: 3,500 RPM for standard tier. Concurrent requests are counted toward RPM. Managing concurrent requests requires queueing. RPM limits are less relevant than TPM for token-heavy applications. Understanding RPM helps with architecture planning.
Last verified: 2026-04-08