Create Response

Call MiniMax models via the OpenAI Responses API compatible main endpoint. Generates model replies, supports streaming and non-streaming.

POST /v1/responses

cURL

SimpleText

curl --request POST \
  --url https://api.minimax.io/v1/responses \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: <content-type>' \
  --data '
{
  "model": "MiniMax-M3",
  "input": "Hello!"
}
'

200

{
  "id": "abc123",
  "object": "response",
  "created_at": 1764000000,
  "model": "MiniMax-M3",
  "status": "completed",
  "output": [
    {
      "id": "abc123_msg",
      "type": "message",
      "status": "completed",
      "role": "assistant",
      "content": [
        {
          "type": "output_text",
          "text": "Hello! I'm MiniMax. How can I help you today?",
          "annotations": []
        }
      ]
    }
  ],
  "output_text": "Hello! I'm MiniMax. How can I help you today?",
  "usage": {
    "input_tokens": 8,
    "input_tokens_details": {
      "cached_tokens": 0
    },
    "output_tokens": 14,
    "output_tokens_details": {
      "reasoning_tokens": 0
    },
    "total_tokens": 22
  },
  "parallel_tool_calls": true,
  "store": false,
  "truncation": "disabled"
}

Reasoning Control

For MiniMax-M3, the reasoning field controls whether the response can include reasoning output.

• If reasoning is omitted, reasoning is disabled by default and the response does not include an output item with type: "reasoning".
• reasoning: {"effort": "none"} is the default behavior and disables reasoning output for MiniMax-M3.
• Values minimal, low, medium, and high are accepted for compatibility and enable reasoning output, but they do not tune MiniMax-M3’s reasoning depth.
• For M2.x models, reasoning cannot be disabled; reasoning: {"effort": "none"} is accepted but reasoning remains on.

{
  "model": "MiniMax-M3",
  "input": "Which is larger, 9.11 or 9.9?"
}

{
  "model": "MiniMax-M3",
  "input": "Which is larger, 9.11 or 9.9?",
  "reasoning": {
    "effort": "minimal"
  }
}

Authorizations

Authorization · string · header · required

HTTP: Bearer Auth

• Security Scheme Type: http
• HTTP Authorization Scheme: Bearer API_key, used to authenticate your account. View it in Account Management > API Keys

Headers

Content-Type · enum<string> · default: application/json · required

Media type of the request body. Must be set to application/json

Available options: application/json

Body

application/json

model · string · required

Model name to invoke, e.g. MiniMax-M3

Example:

"MiniMax-M3"

input · string object[] · required

Conversation content. Supports either a simple text or a full conversation history array

service_tier · enum<string> · default: standard

Service tier for request admission. Supported values are standard and priority. If omitted, the request uses the standard tier. The priority price is 1.5 times the standard price and ensures priority admission so the request is processed ahead of other requests, leading to faster responses and fewer failures.

Available options: standard, priority

instructions · string

System instructions

max_output_tokens · integer

Maximum output token count

temperature · number<float> · default: 1

Sampling temperature, range (0, 1]

Required range: 0 <= x <= 1

top_p · number<float> · default: 0.95

Nucleus sampling, range (0, 1]

Required range: 0 <= x <= 1

stream · boolean · default: false

Set to true to enable SSE streaming response

tools · object[]

Tool list

Show child attributes

tool_choice · enum<string>

Tool selection strategy: none means no tool will be called; auto lets the model decide whether to call tools

Available options: none, auto

metadata · object

Request metadata. Both keys and values are strings

Show child attributes

prompt_cache_key · string

Prompt cache routing identifier

text · object

Output format control

Show child attributes

reasoning · object

Reasoning control. For MiniMax-M3, the default is none, which disables reasoning. Set effort to a non-none value (minimal, low, medium, or high) to enable Adaptive Thinking, but this does not tune MiniMax-M3's reasoning depth. For M2.x models, reasoning cannot be disabled.

Show child attributes

Response

200 - application/json

Successful response

id · string · required

Response ID

Example:

"abc123"

object · enum<string> · required

Object type, always response

Available options: response

created_at · integer · required

Response creation time (Unix seconds)

model · string · required

Actual model that processed the request

status · enum<string> · required

Response status

Available options: completed, incomplete, failed

output · (Message · object | Reasoning · object | Function Call · object)[] · required

Model output list

Assistant reply

Message

Reasoning

Function Call

Show child attributes

output_text · string | null

Convenience field. Concatenation of all text outputs

usage · object

Show child attributes

error · object

Error info, only returned when status=failed

Show child attributes

incomplete_details · object

Reason for incompletion, only returned when status=incomplete

Show child attributes

parallel_tool_calls · boolean

Whether parallel tool calls are supported

store · boolean

Whether the response is persisted

truncation · enum<string>

Context truncation strategy

Available options: disabled