OA
OneAI API
Docs
rateLimitRpmmonthlyBudgetUsdmaxCostUsd

Reference: Rate limits

Quotas, API key limits, and cost controls for production usage.

Best practices

Control spend and protect the gateway before expanding customer limits.

  • Set per-key rateLimitRpm for production keys
  • Set monthlyBudgetUsd where customer spend needs a hard guard
  • Use options.llm.maxCostUsd for expensive task classes
  • Use cheap or balanced modes for default customer traffic
  • Reserve premium and explicit model selection for paid plans

Plan gates

Plans can control task tiers, routing modes, explicit model selection, debug traces, and model registry access.

Plan
Modes
Model selection
Debug / registry
Free
cheap, balanced
Locked
Locked
Pro
cheap, balanced, fast, auto
Locked
Locked
Team
cheap, balanced, fast, premium, auto
Allowed
Allowed
Enterprise
all commercial modes
Allowed
Allowed