RPD
Requests Per Day, a macro-level upper-limit metric for LLM API rate limiting. It typically resets at midnight UTC or Pacific Time.
Timeline (last 90 days)
Rate limiting (429) on free LLM APIs is usually metered simultaneously across four dimensions—RPM, TPM, RPD and concurrency—and different dimensions require different remediation strategies.
Backoff strategies are only effective for short windows such as RPM, TPM and concurrency; for quotas like RPD that reset to zero the same day, backoff is ineffective and one should simply fail the request or switch providers.
Gemini enforces RPM, TPM and daily RPD quotas, and paid tiers also have a rolling 10-minute consumption limit.
Modern LLM rate limiting is a multi-dimensional throughput constraint that evaluates RPM, TPM, RPD and TPD/consumption limits simultaneously, returning HTTP 429 as soon as any dimension hits its ceiling.
All Facts (4)
Rate limiting (429) on free LLM APIs is usually metered simultaneously across four dimensions—RPM, TPM, RPD and concurrency—and different dimensions require different remediation strategies.
50%UnverifiedBackoff strategies are only effective for short windows such as RPM, TPM and concurrency; for quotas like RPD that reset to zero the same day, backoff is ineffective and one should simply fail the request or switch providers.
50%UnverifiedGemini enforces RPM, TPM and daily RPD quotas, and paid tiers also have a rolling 10-minute consumption limit.
50%UnverifiedModern LLM rate limiting is a multi-dimensional throughput constraint that evaluates RPM, TPM, RPD and TPD/consumption limits simultaneously, returning HTTP 429 as soon as any dimension hits its ceiling.
50%