Pay-Per-Token Billing
Self-service AI Gateway access uses pay-per-token billing. Usage is charged for input and output tokens at the published rates below, with no concurrency limit. Each AI Gateway key belongs to one user and one organization. Usage is billed to the organization associated with the key and draws from that organization’s SaladCloud credit balance. The organization must have a positive credit balance when a request is admitted. Follow Pay-Per-Token Onboarding to add credits and create an organization-specific AI Gateway API key. Requests already admitted are allowed to complete and their tokens are billed even if credits run out during generation. New requests are rejected with402 credits_exhausted once the credit balance is no longer positive. Likewise, revoking
or regenerating a key does not cancel or remove charges for requests already admitted. See
key lifecycle behavior.
See Errors for suggested actions and safe retry behavior.
Pay-Per-Token Rates
Prices are per million tokens.
Input tokens include your prompt, image inputs, tool definitions, and any conversation history sent with each request.
Output tokens are the generated response.
Learn more about Salad AI Gateway