OpenAI

POST

/v1/responses

curl --request POST \
  --url https://coder.api.visioncoder.cn/v1/responses \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{ "model": "gpt-5.4", "input": "Introduce VisionCoder in one sentence", "instructions": "example", "reasoning": { "effort": "minimal" }, "max_output_tokens": 4096, "temperature": 1, "top_p": 1, "stream": false, "previous_response_id": "example", "store": true, "tools": [ {} ], "tool_choice": "example", "metadata": { "additionalProperty": "example" } }'

• Production

Authorizations

• BearerAuth

Request Body^required

Media typeapplication/json

OpenAIProvidersReq

OpenAI Responses API request body (recommended for the GPT-5 series). Fields follow the official OpenAI protocol.

object

model

required

OpenAI GPT-5 series model ID, e.g. gpt-5.6-sol, gpt-5.4, gpt-5.4-mini, gpt-5.5, gpt-5.3-codex-spark. Call GET /v1/models for the live list.

string

Example

gpt-5.4

input

required

User input. Either a string or a structured input array

Example

Introduce VisionCoder in one sentence

instructions

System-level instructions, equivalent to the system message in Chat Completions

string

reasoning

object

effort

string

Allowed values: minimal low medium high

Example

high

max_output_tokens

integer

Example

4096

temperature

number format: float

<= 2

top_p

number format: float

stream

boolean

previous_response_id

ID of the previous response, used for multi-turn conversations

string

store

Whether to store the response server-side (mutually exclusive with disable_response_storage)

boolean

tools

Array<object>

object

tool_choice

Tool choice strategy

metadata

object

key

additional properties

string

Responses

200

A successful response.

Media typeapplication/json

OpenAIProvidersResp

OpenAI Responses API response body.

object

id

string

object

string

created_at

integer

status

string

Allowed values: completed in_progress failed cancelled

model

string

output

Structured output containing items such as reasoning / message / tool_call

Array<object>

object

output_text

Concatenated plain-text output (used by some clients)

string

usage

object

input_tokens

integer

output_tokens

integer

total_tokens

integer

Example

{
  "id": "resp_abc123",
  "object": "response",
  "created_at": 1730000000,
  "status": "completed",
  "model": "gpt-5.4",
  "usage": {
    "input_tokens": 12,
    "output_tokens": 28,
    "total_tokens": 40
  }
}