Skip to main content
Rate limits depend on the account and the selected model. View the limits assigned to your account in Higgsfield Console. The primary generation limit is concurrency: the number of requests that may be processing at the same time. Some models can also have model-specific limits.

​
When concurrency is reached

Generation requests above your concurrency limit are still accepted. The API returns a request_id with status queued, and the request starts automatically when one of your running requests reaches a terminal status.

​
When the queue is full

The number of generations waiting to start is limited too. The queue limit is a multiple of your account’s concurrency limit, and each output of a batch request counts separately. When the queue is full, a new generation request is rejected with 429 Too Many Requests and a Retry-After header, and nothing is created or charged:
limit is the queue limit for your account. Wait for Retry-After seconds and submit again with the same Idempotency-Key; queued requests leave the queue as soon as they start. If queue admission is briefly unavailable, the API returns 503 Service Unavailable with "code": "admission_control_unavailable" and "retryable": true. Retry with backoff.

​
Endpoint-specific limits

Some non-generation endpoints reject requests above their limit instead of queuing them. They return 429 Too Many Requests with a Retry-After header:

​
Client recommendations

  • On 429, wait for the number of seconds in Retry-After before submitting again.
  • Limit generation submissions with a worker pool or semaphore so the queue stays short and never reaches its limit.
  • Track each accepted request_id until it becomes terminal.
  • Use backoff and jitter instead of retrying in a tight loop.
  • Keep polling traffic separate from generation submission concurrency.
  • Let us know in Discord before a planned traffic increase if you need higher limits.
Successful responses do not yet include rate-limit headers with your remaining capacity. Treat the limits shown in your dashboard as the authoritative values for your account.