v1

latestOpenAPI 3.1.0Apache 2.02026-07-264001921002.3 KB

OpenAI Integration

Create chat completion (OpenAI format)

Creates a chat completion using OpenAI-compatible format. Supports streaming via SSE.

Async inference: Send x-bf-async: true to submit the request as a background job and receive a job ID immediately. Poll with x-bf-async-id: <job-id> to retrieve the result. When the job is still processing, the response will have an empty choices array. When completed, choices will contain the full result. See Async Inference for details.

Note: This endpoint also works without the /v1 prefix (e.g., /openai/chat/completions).

post/openai/v1/chat/completions

Headers

x-bf-async'true'

Set to true to submit this request as an async job. Returns immediately with a job ID. Not compatible with streaming.

x-bf-async-idstring

Poll for results of a previously submitted async job by providing the job ID returned from the initial async request.

x-bf-async-job-result-ttlinteger

Override the default result TTL in seconds. Results expire after this duration from completion time.

Request body

modelstring required

Model identifier (e.g., gpt-4, gpt-3.5-turbo)

streamboolean

Whether to stream the response

max_tokensinteger

Maximum tokens to generate (legacy, use max_completion_tokens)

max_completion_tokensinteger

Maximum tokens to generate

temperaturenumber

top_pnumber

frequency_penaltynumber

presence_penaltynumber

logit_biasobject

logprobsboolean

top_logprobsinteger

ninteger

seedinteger

userstring

parallel_tool_callsboolean

response_formatobject

Format for the response

OpenAI reasoning effort level

service_tierstring

fallbacksstring[]

Fallback models

Example request

{
  "model": "gpt-4"
}

Response

Successful response

idstring

createdinteger

modelstring

objectstring

service_tierstring

system_fingerprintstring

citationsstring[]