ResponsesRequest

Request body for OpenAI Responses API.

This is the API wire type for the Responses endpoint. The canonical internal task params type is TextGenerationTaskParams; see the responses_request_to_text_generation adapter for conversion.

modelModel (string)required

input objectrequired

anyOf

• string
• object[]

string

instructions object

anyOf

• string
• null

string

max_output_tokens object

anyOf

• integer
• null

integer

temperature object

anyOf

• number
• null

number

top_p object

anyOf

• number
• null

number

streamStream (boolean)

Default value: false

tools object

anyOf

• object[]
• null

Array [

property name*any

]

metadata object

anyOf

• object
• null

object

reasoning object

anyOf

• Reasoning
• null

effort object

anyOf

• string
• null

string

Possible values: [none, minimal, low, medium, high, xhigh]

summary object

anyOf

• string
• null

string

Possible values: [auto, concise, detailed]

enable_thinking object

[Skulk extension] Boolean thinking toggle. Not part of the OpenAI Responses API.

anyOf

• boolean
• null

boolean

top_k object

[Skulk extension] Top-k sampling parameter. Not part of the OpenAI Responses API.

anyOf

• integer
• null

integer

stop object

[Skulk extension] Stop sequence(s). Not part of the OpenAI Responses API.

anyOf

• string
• string[]
• null

string

seed object

[Skulk extension] Seed for deterministic sampling. Not part of the OpenAI Responses API.

anyOf

• integer
• null

integer

chat_template_messages object

Internal: pre-formatted messages for tokenizer chat template. Not part of the OpenAI Responses API.

anyOf

• object[]
• null

Array [

property name*any

]

ResponsesRequest

{
  "input": "Hello from the Responses API",
  "model": "mlx-community/Llama-3.2-1B-Instruct-4bit",
  "stream": false
}