MaaS_MiniMax_M2.5/2.7

Model Name MiniMax-M2.5/M2.7

Context Window 204,800

Model introduction, top performance and ultimate cost-effectiveness, easily handle complex tasks (output speed is about 60 TPS).

Model Name MiniMax-M2.5-highspeed/MiniMax-M2.7-highspeed

Context Window 204,800

Model Introduction M2.5 Extreme Edition: The effect remains unchanged, but it is faster and more agile (output speed is approximately 100 TPS)

Request Protocol

Https

Request Header

Parameter Name	Value
Authorization	Bearer
Content-Type	multipart/form-data

Parameter

Core Parameters

Attribute Name	Type	Required/Optional	Description
model	string	Required	Model Name.
messages	array	Required	Conversation message list, containing conversation history. Supports basic roles (system, user, assistant) and advanced roles (user_system, group, sample_message_user, sample_message_ai).
temperature	number	Optional	Temperature coefficient, controls output randomness, with a value range of (0, 1]. The higher the value, the more random the output; the lower the value, the more deterministic the output, with a default value of 1.0.
top_p	number	Optional	Nucleus sampling strategy, which controls output diversity, with a value range of (0, 1]. The smaller the value, the more conservative the output.
max_tokens / max_completion_tokens	integer	Optional	The maximum number of tokens for generated content, with an upper limit of 2048. It is recommended to set this value reasonably to avoid output truncation.
stream	boolean	Optional	Whether to use streaming, default is false. Setting it to true will return the response in batches.

Advanced Function Parameters

Attribute Name	Type	Required/Optional	Description
stop	string/array	Optional	Up to 4 stop sequences, upon encountering which the API will stop generating tokens.
n	integer	Optional	Number of chat completion options generated for each input message. MiniMax currently only supports n = 1.
frequency_penalty	number	Optional	Frequency Penalty, with a value range from -2.0 to 2.0. A positive value penalizes tokens that have already appeared frequently, reducing the probability of the model repeating statements.
presence_penalty	number	Optional	There is a penalty, with a value range from -2.0 to 2.0. Positive values penalize tokens that have already appeared, encouraging the model to explore new topics.
response_format	object	Optional	Specify the model output format. Setting {"type": "json_object"} enables JSON mode.
seed	integer	Optional	Seed for deterministic sampling. Under the same seed and parameters, the system will strive to return the same results.
tool_choice	string/object	Optional	Controls whether the model calls tools. Optional values: "none" (no call), "auto" (automatic selection), "required" (must call), or specify a specific function.
tools	array	Optional	The list of tools that the model can call, currently only supports function type.
mask_sensitive_info	boolean	Optional	Whether to desensitize sensitive information in the output, default is true.
user	string	Optional	Unique identifier of the end user, used for monitoring and abuse detection.

Special role parameters (available in the messages array)

Attribute Name	Type	Required/Optional	Description
role	string	Required	Role types. Basic roles: system (sets the model role), user (user input), assistant (model response). Advanced roles: user_system (sets the user persona), group (conversation grouping), sample_message_user (example user input), sample_message_ai (example model output).
name	string	Optional	The name identifier of the message sender, used to distinguish different participants in the conversation.
content	string	Condition Required	The text content of the message. It is a required field for regular conversation messages and for sample conversations (sample_message_user/sample_message_ai).

/chat/completions Request Example

curl --location --request POST 'https://genaiapi.cloudsway.net/v1/ai/{endpointPath}/chat/completions' \
--header 'Authorization: Bearer {YOUR AK}' \
--header 'Content-Type: application/json' \
--data-raw '{
    "model": "MiniMax-M2.5",
    "reasoning_split": true,
    "messages": [
      {
        "role": "user",
        "content": [
          {
            "type": "text",
            "text": "what is elephent？"
          }
        ]
      }
    ]
  }'

Return Example

{
    "id": "chatcmpl-wY8FIAgqKvdcHBAZsSL3n2H1",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "Elephants are mammals belonging to the family Elephantidae. They are the largest land animals, mainly divided into two genera: Loxodonta (African elephants) and Elephas (Asian elephants). Elephants are known for their enormous size, long trunk, and large ears. They live in a variety of habitats such as grasslands, forests, and swamps, and are herbivores, feeding on grasses, leaves, bark, fruits, and more. Elephants are highly social animals, typically living in groups, and have complex social structures and communication methods.
",
                "reasoning_details": [
                    {
                        "type": "reasoning.text",
                        "text": "<answer>\nElephants are mammals belonging to the family Elephantidae. They are the largest land animals, mainly divided into two genera: Loxodonta (African elephants) and Elephas (Asian elephants). Elephants are known for their enormous size, long trunk, and large ears. They live in a variety of habitats such as grasslands, forests, and swamps, and are herbivores, feeding on grasses, leaves, bark, fruits, and more. Elephants are highly social animals, typically living in groups, and have complex social structures and communication methods\n</answer>\n",
                        "signature": null
                    }
                ]
            },
            "finish_reason": "stop",
            "native_finish_reason": "stop"
        }
    ],
    "created": 1779174874,
    "model": "MaaS_Minimax_M2.7_20260318",
    "object": "chat.completion",
    "usage": {
        "prompt_tokens": 46,
        "completion_tokens": 203,
        "total_tokens": 249,
        "completion_tokens_details": {
            "accepted_prediction_tokens": 0,
            "audio_tokens": 0,
            "image_tokens": 0,
            "reasoning_tokens": 172,
            "rejected_prediction_tokens": 0
        },
        "prompt_tokens_details": {
            "audio_tokens": 0,
            "cached_tokens": 35,
            "image_tokens": 0
        }
    }
}

Text Conversation (OpenAI API Compatible)

Request Example

curl --request POST \
  --url https://{新平台域名}/v1/chat/completions \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "MaaS_Minimax_M2.7_20260318",
  "messages": [
    {
      "role": "system",
      "name": "MiniMax AI"
    },
    {
      "role": "user",
      "content": "hello",
      "name": "user"
    }
  ]
}
'

Return Example

{
    "id": "chatcmpl-Sfq4bWiUnaarkHytvNwBcrRw",
    "choices": [
        {
            "index": 0,
            "message": {
                "role": "assistant",
                "content": "Hello! Is there anything I can help you with?"
            },
            "finish_reason": "stop",
            "native_finish_reason": "stop"
        }
    ],
    "created": 1779175080,
    "model": "MaaS_Minimax_M2.7_20260318",
    "object": "chat.completion",
    "usage": {
        "prompt_tokens": 43,
        "completion_tokens": 29,
        "total_tokens": 72,
        "completion_tokens_details": {
            "accepted_prediction_tokens": 0,
            "audio_tokens": 0,
            "image_tokens": 0,
            "reasoning_tokens": 33,
            "rejected_prediction_tokens": 0
        },
        "prompt_tokens_details": {
            "audio_tokens": 0,
            "cached_tokens": 35,
            "image_tokens": 0
        }
    }
}