Responses API | Tutujin API Docs

POST /v1/responses is OpenAI's newer API format: input replaces messages, system instructions go in instructions, and reasoning effort, structured output and tools are written differently from Chat Completions. Some models in the catalog only support this format. Requests are relayed to the model service in their original format, so whether a parameter takes effect follows that model's official documentation.

POSThttps://open.tutujin.com/v1/responses

Headers

Header	Required	Description
`Authorization`	Yes	`Bearer sk-...`, an API key from the console
`Content-Type`	Yes	`application/json`

Request body

Parameter	Type	Required	Description
`model`	string	Yes	Model name, e.g. `gpt-6.1-sol`
`input`	string or array	Yes	The input. A plain string, or an array of messages for multi-turn chats and image input — structure below
`instructions`	string	No	System instructions, like the `system` message in Chat Completions
`max_output_tokens`	integer	No	Most tokens to generate, including reasoning tokens
`temperature`	number	No	Sampling temperature; higher is more varied. Reasoning models usually do not support it
`top_p`	number	No	Nucleus sampling; tune this or `temperature`, not both
`stream`	boolean	No	`true` returns server-sent events; default `false`
`reasoning`	object	No	Reasoning settings, e.g. `{"effort": "low"}` with `low`, `medium` or `high`; reasoning models only
`text`	object	No	Output format, e.g. `{"format": {"type": "json_object"}}` for JSON; some models support `json_schema` for strict structured output
`tools`	array	No	Tools the model may call, such as function definitions. Built-in tools depend on the model
`tool_choice`	string or object	No	`auto`, `none`, `required`, or a specific tool
`store`	boolean	No	Whether the response is stored server-side. Features that rely on stored responses are not verified on this site; set it to `false`
`previous_response_id`	string	No	Continues from a stored response. It relies on server-side storage and is not verified here, so send the history in `input` instead
`metadata`	object	No	Your own key-value pairs for reconciliation and debugging

The input field

The simplest form is a string. For multi-turn chats or images, pass an array of messages:

JSON

[
  { "role": "user", "content": "What is the total on this invoice?" },
  {
    "role": "user",
    "content": [
      { "type": "input_text", "text": "Reply with the amount and currency only." },
      { "type": "input_image", "image_url": "https://example.com/invoice.png" }
    ]
  }
]

role is user, assistant, system or developer. You can also inline an image in image_url as data:image/png;base64,....

Request example

curl https://open.tutujin.com/v1/responses \
  -H "Authorization: Bearer $TUTUJIN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6.1-sol",
    "instructions": "You are a finance assistant. Be brief and use two decimal places.",
    "input": "Travel cost 1280, meal allowance 80 a day for 3 days. What is the total claim?"
  }'

import os
from openai import OpenAI

client = OpenAI(base_url="https://open.tutujin.com/v1", api_key=os.environ["TUTUJIN_API_KEY"])

resp = client.responses.create(
    model="gpt-6.1-sol",
    instructions="You are a finance assistant. Be brief and use two decimal places.",
    input="Travel cost 1280, meal allowance 80 a day for 3 days. What is the total claim?",
)
print(resp.output_text)
print(resp.usage)

import OpenAI from 'openai'

const client = new OpenAI({ baseURL: 'https://open.tutujin.com/v1', apiKey: process.env.TUTUJIN_API_KEY })

const resp = await client.responses.create({
  model: 'gpt-6.1-sol',
  instructions: 'You are a finance assistant. Be brief and use two decimal places.',
  input: 'Travel cost 1280, meal allowance 80 a day for 3 days. What is the total claim?',
})
console.log(resp.output_text)
console.log(resp.usage)

Response example

JSON

{
  "id": "resp_...",
  "object": "response",
  "created_at": 1794000000,
  "status": "completed",
  "model": "gpt-6.1-sol",
  "output": [
    {
      "type": "message",
      "id": "msg_...",
      "status": "completed",
      "role": "assistant",
      "content": [
        { "type": "output_text", "text": "The total claim is 1520.00.", "annotations": [] }
      ]
    }
  ],
  "usage": {
    "input_tokens": 48,
    "input_tokens_details": { "cached_tokens": 0 },
    "output_tokens": 15,
    "output_tokens_details": { "reasoning_tokens": 0 },
    "total_tokens": 63
  }
}

Field	Description
`status`	`completed` finished normally; `incomplete` usually means `max_output_tokens` cut it off
`output`	Output items: `message` (the answer), `reasoning` (a reasoning summary), `function_call` (a function for you to run)
`output[].content[].text`	The answer text. The official SDKs expose `output_text` with all text joined
`usage.input_tokens`	Input tokens, billed at the input price
`usage.input_tokens_details.cached_tokens`	Input tokens served from the context cache, billed at the cache-read price
`usage.output_tokens`	Output tokens, billed at the output price; `reasoning_tokens` is the thinking part

Streaming

With "stream": true the server pushes server-sent events. Each event has an event: name; new text arrives in the delta field of response.output_text.delta, and response.completed carries the full response and usage:

Text

event: response.created
data: {"type":"response.created","response":{"id":"resp_...","status":"in_progress"}}

event: response.output_text.delta
data: {"type":"response.output_text.delta","output_index":0,"content_index":0,"delta":"The total"}

event: response.output_text.done
data: {"type":"response.output_text.done","output_index":0,"content_index":0,"text":"The total claim is 1520.00."}

event: response.completed
data: {"type":"response.completed","response":{"id":"resp_...","status":"completed","usage":{"input_tokens":48,"output_tokens":15,"total_tokens":63}}}

With an SDK, pick text out by event type:

stream = client.responses.create(
    model="gpt-6.1-sol",
    input="Write a 60-word product description",
    stream=True,
)
for event in stream:
    if event.type == "response.output_text.delta":
        print(event.delta, end="", flush=True)
    elif event.type == "response.completed":
        print("\n", event.response.usage)

const stream = await client.responses.create({
  model: 'gpt-6.1-sol',
  input: 'Write a 60-word product description',
  stream: true,
})
for await (const event of stream) {
  if (event.type === 'response.output_text.delta') process.stdout.write(event.delta)
  if (event.type === 'response.completed') console.log('\n', event.response.usage)
}

Notes

This endpoint is relayed as-is and has not been verified separately on this site. Check parameters and response shape with a little traffic when you adopt a new model.
For multi-turn chats, send the full history in input rather than relying on previous_response_id and store.
Reasoning tokens count as output, so reasoning models can cost noticeably more; cap them with max_output_tokens and reasoning.effort.
Whether built-in tools such as web search work, and whether they are billed separately, depends on the model; the Usage page in the console is the record.
If a model also supports Chat Completions and you do not need Responses-only features, Chat Completions is simpler, and switching vendors only means changing model.

Models that support this endpoint

Model	Vendor	Cheapest group	Input ¥/1M	Output ¥/1M	Per request ¥
GPT-6.1 Sol	OpenAI	Standard	¥4	¥20	—
GPT-6 Luna	OpenAI	Standard	¥0.2	¥1	—
GPT-6 Sol	OpenAI	Standard	¥20	¥100	—
gpt-6-astra	OpenAI	Standard	¥20	¥100	—
muse-wonder-1.0	Other	Standard	¥56	¥280	—

Prices in CNY, cheapest group for each model; updated Nov 12, 2026, 8:17 PM.

Still stuck? Add our WeChat support JsnonoChat and send the request id from the error message, and we will look into it.

POSThttps://open.tutujin.com/v1/responses

Headers

Header	Required	Description
`Authorization`	Yes	`Bearer sk-...`, an API key from the console
`Content-Type`	Yes	`application/json`

Request body

Parameter	Type	Required	Description
`model`	string	Yes	Model name, e.g. `gpt-6.1-sol`
`input`	string or array	Yes	The input. A plain string, or an array of messages for multi-turn chats and image input — structure below
`instructions`	string	No	System instructions, like the `system` message in Chat Completions
`max_output_tokens`	integer	No	Most tokens to generate, including reasoning tokens
`temperature`	number	No	Sampling temperature; higher is more varied. Reasoning models usually do not support it
`top_p`	number	No	Nucleus sampling; tune this or `temperature`, not both
`stream`	boolean	No	`true` returns server-sent events; default `false`
`reasoning`	object	No	Reasoning settings, e.g. `{"effort": "low"}` with `low`, `medium` or `high`; reasoning models only
`text`	object	No	Output format, e.g. `{"format": {"type": "json_object"}}` for JSON; some models support `json_schema` for strict structured output
`tools`	array	No	Tools the model may call, such as function definitions. Built-in tools depend on the model
`tool_choice`	string or object	No	`auto`, `none`, `required`, or a specific tool
`store`	boolean	No	Whether the response is stored server-side. Features that rely on stored responses are not verified on this site; set it to `false`
`previous_response_id`	string	No	Continues from a stored response. It relies on server-side storage and is not verified here, so send the history in `input` instead
`metadata`	object	No	Your own key-value pairs for reconciliation and debugging

The input field

The simplest form is a string. For multi-turn chats or images, pass an array of messages:

JSON

[
  { "role": "user", "content": "What is the total on this invoice?" },
  {
    "role": "user",
    "content": [
      { "type": "input_text", "text": "Reply with the amount and currency only." },
      { "type": "input_image", "image_url": "https://example.com/invoice.png" }
    ]
  }
]

role is user, assistant, system or developer. You can also inline an image in image_url as data:image/png;base64,....

Request example

curl https://open.tutujin.com/v1/responses \
  -H "Authorization: Bearer $TUTUJIN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-6.1-sol",
    "instructions": "You are a finance assistant. Be brief and use two decimal places.",
    "input": "Travel cost 1280, meal allowance 80 a day for 3 days. What is the total claim?"
  }'

import os
from openai import OpenAI

client = OpenAI(base_url="https://open.tutujin.com/v1", api_key=os.environ["TUTUJIN_API_KEY"])

resp = client.responses.create(
    model="gpt-6.1-sol",
    instructions="You are a finance assistant. Be brief and use two decimal places.",
    input="Travel cost 1280, meal allowance 80 a day for 3 days. What is the total claim?",
)
print(resp.output_text)
print(resp.usage)

import OpenAI from 'openai'

const client = new OpenAI({ baseURL: 'https://open.tutujin.com/v1', apiKey: process.env.TUTUJIN_API_KEY })

const resp = await client.responses.create({
  model: 'gpt-6.1-sol',
  instructions: 'You are a finance assistant. Be brief and use two decimal places.',
  input: 'Travel cost 1280, meal allowance 80 a day for 3 days. What is the total claim?',
})
console.log(resp.output_text)
console.log(resp.usage)

Response example

JSON

{
  "id": "resp_...",
  "object": "response",
  "created_at": 1794000000,
  "status": "completed",
  "model": "gpt-6.1-sol",
  "output": [
    {
      "type": "message",
      "id": "msg_...",
      "status": "completed",
      "role": "assistant",
      "content": [
        { "type": "output_text", "text": "The total claim is 1520.00.", "annotations": [] }
      ]
    }
  ],
  "usage": {
    "input_tokens": 48,
    "input_tokens_details": { "cached_tokens": 0 },
    "output_tokens": 15,
    "output_tokens_details": { "reasoning_tokens": 0 },
    "total_tokens": 63
  }
}

Field	Description
`status`	`completed` finished normally; `incomplete` usually means `max_output_tokens` cut it off
`output`	Output items: `message` (the answer), `reasoning` (a reasoning summary), `function_call` (a function for you to run)
`output[].content[].text`	The answer text. The official SDKs expose `output_text` with all text joined
`usage.input_tokens`	Input tokens, billed at the input price
`usage.input_tokens_details.cached_tokens`	Input tokens served from the context cache, billed at the cache-read price
`usage.output_tokens`	Output tokens, billed at the output price; `reasoning_tokens` is the thinking part

Streaming

Text

event: response.created
data: {"type":"response.created","response":{"id":"resp_...","status":"in_progress"}}

event: response.output_text.delta
data: {"type":"response.output_text.delta","output_index":0,"content_index":0,"delta":"The total"}

event: response.output_text.done
data: {"type":"response.output_text.done","output_index":0,"content_index":0,"text":"The total claim is 1520.00."}

event: response.completed
data: {"type":"response.completed","response":{"id":"resp_...","status":"completed","usage":{"input_tokens":48,"output_tokens":15,"total_tokens":63}}}

With an SDK, pick text out by event type:

stream = client.responses.create(
    model="gpt-6.1-sol",
    input="Write a 60-word product description",
    stream=True,
)
for event in stream:
    if event.type == "response.output_text.delta":
        print(event.delta, end="", flush=True)
    elif event.type == "response.completed":
        print("\n", event.response.usage)

const stream = await client.responses.create({
  model: 'gpt-6.1-sol',
  input: 'Write a 60-word product description',
  stream: true,
})
for await (const event of stream) {
  if (event.type === 'response.output_text.delta') process.stdout.write(event.delta)
  if (event.type === 'response.completed') console.log('\n', event.response.usage)
}

Notes

This endpoint is relayed as-is and has not been verified separately on this site. Check parameters and response shape with a little traffic when you adopt a new model.
For multi-turn chats, send the full history in input rather than relying on previous_response_id and store.
Reasoning tokens count as output, so reasoning models can cost noticeably more; cap them with max_output_tokens and reasoning.effort.
Whether built-in tools such as web search work, and whether they are billed separately, depends on the model; the Usage page in the console is the record.
If a model also supports Chat Completions and you do not need Responses-only features, Chat Completions is simpler, and switching vendors only means changing model.

Models that support this endpoint

Model	Vendor	Cheapest group	Input ¥/1M	Output ¥/1M	Per request ¥
GPT-6.1 Sol	OpenAI	Standard	¥4	¥20	—
GPT-6 Luna	OpenAI	Standard	¥0.2	¥1	—
GPT-6 Sol	OpenAI	Standard	¥20	¥100	—
gpt-6-astra	OpenAI	Standard	¥20	¥100	—
muse-wonder-1.0	Other	Standard	¥56	¥280	—

Prices in CNY, cheapest group for each model; updated Nov 12, 2026, 8:17 PM.