Skip to main content
Each API key has its own per-minute budgets, split by what the request does: A rate-limited request returns 429 Too Many Requests with a Retry-After header in seconds. Wait that long, then send the same request again. Reads and status polling share the GET budget, which is separate from the generation budget, so polling a running job never uses up the budget for starting new ones. Keep a few seconds between polls; see Polling.
Last modified on September 30, 2026