Service tiers now available on AI Gateway
By Steven Van ·
Developers can request a faster priority tier or a cheaper flex tier for OpenAI and Gemini calls, across every AI Gateway API format.
Vercel's AI Gateway now supports service tiers for OpenAI and Gemini models, letting a request trade off speed and cost. The priority tier processes faster at a higher cost, the flex tier costs less but can run slower, and requests that don't specify a tier fall back to the default tier at standard pricing.
Service tier is set with a single serviceTier option under providerOptions.gateway, so switching models or providers doesn't require restructuring the request, and it works the same way across the AI SDK, Chat Completions API, Anthropic Messages API, OpenAI Responses API, and OpenResponses API. Gateway bills each request at whichever tier the provider actually served rather than the tier requested, and returns the applied tier in the response's provider metadata. If a provider can't honour the requested tier, the request still runs at the default rate; only an invalid tier value causes a failure. Read the announcement for setup details.