Anthropic-compatible APIs

Constraints

This function is only supported in CN-Hong Kong.

API Information

Parameter	Description	Example Value
API URL	API URL for calling the model service.	https://api-ap-southeast-1.modelarts-maas.com/anthropic/v1/messages
Authentication request header	Provides the server with identity authentication/authorization credentials.	"x-api-key":"$MaaS_API_Key"

Supported Models

You can log in to the MaaS console and enable the following models on the Model Inference > Real-Time Inference > Built-in Services page. On the model details page in the Model Square, you can view detailed information about the models.

Model	Version	Supported Region	Value of model
GLM	GLM-5	CN-Hong Kong	glm-5
	GLM-5.1	CN-Hong Kong	glm-5.1
	GLM-5.2	CN-Hong Kong	glm-5.2
DeepSeek	DeepSeek-V4-Pro	CN-Hong Kong	deepseek-v4-pro
	DeepSeek-V4-Flash	CN-Hong Kong	deepseek-v4-flash
	DeepSeek-V3	CN-Hong Kong	DeepSeek-V3
	DeepSeek-V3.2	CN-Hong Kong	deepseek-v3.2
	DeepSeek-R1-0528	CN-Hong Kong	deepseek-r1-250528

Request Parameters

Category	Parameter	Type	Mandatory	Description
General	model	string	Yes	Definition Model parameter used for generating content. For details about its value, see the model values in Supported models. Constraints N/A Range For details about its value, see the model values in Supported models. Default Value N/A
	system	string	No	Definition System prompt, which can be used to assign a specific goal or role to the model. Constraints N/A Range N/A Default Value N/A
	messages	object[]	Yes	Definition List of messages in a dialog. Different models support different types of messages.
	messages.role	enum<string>	Yes	Definition Role that sends a message. Constraints N/A Range user assistant: A message returned by the model. Default Value N/A
	messages.content	string/array	Yes	Definition Message sent by the current role. Constraints N/A Range N/A Default Value N/A
	messages.content.type	enum<string>	Yes	Definition Message type. Constraints N/A Range text: text messages. tool_use: tool calls. tool_result: tool call results. Default Value N/A
Tool call result	messages.content.tool_use_id	string	Yes	Definition ID of the tool called by the model. Constraints This parameter is mandatory when messages.content.type is set to tool_result. Range N/A Default Value N/A
Tool call result	messages.content.content	string	No	Definition Content obtained after the tool is called. Constraints This parameter is mandatory when messages.content.type is set to tool_result. Range N/A Default Value N/A
Deep thinking control	thinking	object	No	Definition Controls whether to enable the deep thinking mode for a model. Constraints For details about the supported models, see Thinking Control for Deep Thinking Models.
Deep thinking control	thinking.type	string	Yes	Definition Controls whether to enable the deep thinking mode for a model. Constraints For details about the supported models, see Thinking Control for Deep Thinking Models. Range enabled: The deep thinking mode is enabled. The model must think before answering. disabled: The deep thinking mode is disabled. The model directly answers questions without thinking. Default Value The default value varies depending on the model. For details, see Thinking Control for Deep Thinking Models.
Model output behavior configuration	max_tokens	integer/null	No	Definition Maximum length of the content part of the model answer. After this parameter is set, the token length of the model output content does not exceed the value of max_tokens. Constraints N/A Range The default value varies depending on the model. For details, see the maximum output length and maximum CoT length on the model details page. Default Value 4K
	temperature	number	No	Definition Sampling temperature, used to control the randomness of the content generated by the model. A higher value (for example, greater than 1) will make the model output more random, divergent, and creative, while a lower value (for example, 0.1) will make the output more focused, deterministic, and conservative. Constraints It is recommended that you adjust either temperature or top_p. Range [0,2] Default Value DeepSeek-V3, DeepSeek-R1-0528, and DeepSeek-V4 apply special handling to inputs: DeepSeek-V3: If the temperature is less than or equal to 1, it is multiplied by 0.3. If the temperature is greater than 1, 0.7 is subtracted from it. If the temperature is empty, it defaults to 0.6. DeepSeek-R1-0528: Regardless of the user input, the value remains 0.6. DeepSeek-V4: In thinking mode, the value is always 1.0, regardless of the user input. For other models, the input is used if provided; otherwise, it defaults to 0.6.
	top_p	float/null	No	Definition Nucleus sampling probability threshold, used to control the diversity of the content generated by the model. Similar to the temperature parameter, it influences randomness, but offers finer control by dynamically adjusting the pool of possible tokens. When set close to 0, the model samples from only the most probable tokens, resulting in very conservative and highly deterministic outputs. When set close to 1, there is virtually no restriction on the vocabulary, leading to more random and divergent outputs. Constraints It is recommended that you adjust either temperature or top_p. Range (0,1] Default Value DeepSeek-V4: In thinking mode, the value is always 1.0, regardless of the user input. This parameter is not applicable to other models.
	stream	boolean/null	No	Definition Determines whether the response is returned as a stream. Constraints N/A Range false: Returns results once all content has been generated by the model. true: Returns content piece by piece, ending with data: [DONE]. Default Value false
	stop_sequences	string[]	No	Definition Stops the model from generating content when it encounters the string specified in the stop_sequences field. The stop token does not hit the token in the CoT, but only hits the part in the formal answer content. Constraints N/A Range N/A Default Value N/A
	reasoning_effort	string	No	Definition Controls the reasoning intensity. Constraints This field takes effect only in thinking mode. Involved models: DeepSeek-V4 series and GLM-5.2 Range GLM-5.2: max or xhigh: deep reasoning high, medium, or low: enhanced reasoning minimal or none: The model gives up thinking. NOTE: xhigh is mapped to max. low or medium is mapped to high. DeepSeek-V4 series: high Default Value GLM-5.2: max DeepSeek-V4 series: high
Tool calling	tools	array	No	Definition List of tools available for the model to call, which may be included in the model's response. Configure this structure when you require the model to specify the tools to be called. Constraints N/A
	tools.name	string	Yes	Definition Name of the function to be called. Constraints N/A Range N/A Default Value N/A
	tools.description	string	No	Definition Description of the function to be called. The model will determine if the tool is needed for the current task based on this description. Constraints N/A Range N/A Default Value N/A
	tools.input_schema	object	No	Definition Function request parameters, described in JSON format. Example: { "type": "object", "properties": { "parameter name": { "type": "parameter type", "description": "parameter description" } }, "required": ["mandatory parameters"] } Parameters must conform to JSON specifications. Constraints N/A

Response Parameters (Non-Streaming)

Parameter		Type	Description
id		string	Definition Unique identifier of this request. Range N/A
type		enum<string>	Definition The type of the response message. Range message
role		string	Definition The role of the author of this message. Range assistant: A message returned by the model.
model		string	Definition Model parameter used in this request. Range Model parameter used in this request.
content		array	Definition Content generated by the model.
content.type		enum<string>	Definition Type of the generated content. Range text: text messages. thinking: deep thinking. tool_use: tool calls. tool_result: tool call results.
Deep thinking response (type: thinking)	content.thinking	string	Definition Deep thinking content of the model. Range N/A
Deep thinking response (type: thinking)	content.signature	string	Definition Complete encrypted thinking content. Range N/A
Formal reply (type: text)	content.text	string	Definition Formal reply content of the model. Range N/A
Tool input response (type: tool_use)	content.id	string	Definition ID of the tool called. Range N/A
	content.name	string	Definition Name of the tool called. Range N/A
	content.input	object	Definition Structured input required to call the tool. Range N/A
stop_reason	enum<string>/null		Definition Reason why the model stops generating tokens. Range end_turn: The model output ends naturally. max_tokens: The model output triggers the max_tokens limit. stop_sequence: The model output token matches the token in stop_sequence. tool_use: The model determines that a tool needs to be called.
stop_sequence	string/null		Definition Indicates which custom stop_sequence was matched. The model returns null if none were matched. Range N/A
usage	object		Definition Number of tokens consumed by this request. Range N/A
	usage.output_tokens	integer	Definition The number of tokens consumed by the model to generate the completion response. Range N/A
	usage.input_tokens	integer	Definition The number of tokens consumed by the model to process the input. Range N/A
	usage.cache_creation_input_tokens	integer	Definition Number of input tokens newly written to the cache in this request. Range N/A
	usage.cache_read_input_tokens	integer	Definition Number of input tokens served from the cache in this request. Range N/A
	usage.service_tier	string	Definition Service tier used for this request. Range default
	usage.total_tokens	integer	Definition The total number of tokens consumed by the model to process the input and generate the output. Range N/A

Response Parameters (Streaming)

Parameter		Type	Description
event		-	Definition ID of a returned message. Range message_start: the first chunk. content_block_start: the start chunk of the formal reply. content_block_delta: contains the text of the formal reply, input_json for tool calls, and thinking for deep thinking. content_block_stop: indicates that a segment of content has been fully output. message_delta: contains stop_reason, stop_sequence, and usage information. message_stop: indicates that all messages have been returned.
event: message_start	type	string	Definition Type of the chunk. Range message_start: the first chunk.
	message	object[]	Definition Response message.
	message.id	string	Definition Unique ID of a request. Range N/A
	message.type	string	Definition The type of the response message. Range message
	message.role	string	Definition The role of the information. Range assistant
	message.model	string	Definition Model parameter called. Range Model parameter used in this request.
	message.stop_sequence	string/null	Definition Hit stop sequence. Range If no match is found in the cache, the output is null.
	message.usage	object	Definition Token usage statistics of the current request.
	message.usage.input_tokens	integer	Definition The number of tokens consumed by the model to process the input. Range N/A
	message.usage.output_tokens	integer	Definition The number of tokens consumed by the model to generate the completion response. Range N/A
	message.usage.cache_creation_input_tokens	integer	Definition Number of input tokens newly written to the cache in this request. Range N/A
	message.usage.cache_read_input_tokens	integer	Definition Number of input tokens served from the cache in this request. Range N/A
	message.usage.service_tier	string	Definition Service tier used for this request. Range default
	message.usage.total_tokens	integer	Definition The total number of tokens consumed by the model to process the input and generate the output. Range N/A
	message.content	array	Definition Content generated by the model.
	message.stop_reason	enum<string>/null	Definition Reason why the model stops outputting. Range If the model does not stop outputting, the value is null.
event: content_block_start	type	string	Definition Type of the chunk. Range content_block_start: the start chunk of the formal reply.
	index	integer	Definition Index of the current element. Range N/A
	content_block	object	Definition Details about the content generated by the model.
	content_block.type	string	Definition Type of the content generated by the model. Range text
	content_block.thinking	string	Definition Deep thinking content of the model. Range N/A
	content_block.text	string	Definition Formal reply content of the model. Range N/A
	content_block.id	string	Definition ID of the tool called. Range N/A
	content_block.name	string	Definition Name of the tool called. Range N/A
	content_block.input	object	Definition Structured input required to call the tool. Range N/A
event: content_block_delta	type	string	Definition Type of the chunk. Range content_block_delta: contains the text of the formal reply, input_json for tool calls, and thinking for deep thinking.
	index	integer	Definition Index of the current element. Range N/A
	delta	object	Definition Incremental output from the model.
	delta.type	string	Definition Incremental content type. Range text_delta
	delta.text	string	Definition Incremental text output from the model. Range N/A
	delta.partial_json	string	Definition JSON data required for tool calls during streaming output.
event: content_block_stop	type	string	Definition Type of the chunk. Range content_block_stop: indicates that a segment of content has been fully output.
event: content_block_stop	index	integer	Definition Index of the current element. Range N/A
event: message_delta	type	string	Definition Type of the chunk. Range message_delta: contains stop_reason, stop_sequence, and usage information.
	delta	object	Definition Incremental content of the message.
	delta.stop_sequence	string/null	Definition Hit stop sequence. Range If no match is found in the cache, the model returns null.
	delta.stop_reason	string	Definition Reason why the model stops outputting. Range end_turn: The model output ends naturally. max_tokens: The model output triggers the max_tokens limit. stop_sequence: The model output token matches the token in stop_sequence. tool_use: The model determines that a tool needs to be called.
	usage	enum<string>/null	Definition Token usage statistics of the current request.
	usage.output_tokens	integer	Definition The number of tokens consumed by the model to generate the completion response. Range N/A
	usage.input_tokens	integer	Definition The number of tokens consumed by the model to process the input. Range N/A
	usage.cache_creation_input_tokens	integer	Definition Number of input tokens newly written to the cache in this request. Range N/A
	usage.cache_read_input_tokens	integer	Definition Number of input tokens served from the cache in this request. Range N/A
	usage.total_tokens	integer	Definition The total number of tokens consumed by the model to process the input and generate the output. Range N/A
event: message_stop	type	string	Definition Type of the chunk. Range message_stop: indicates that all messages have been returned.

Thinking Control for Deep Thinking Models

Model	thinking.type Default Value	thinking.type Options
GLM-5.2 GLM-5.1 GLM-5	enabled	enabled disabled
DeepSeek-V4-Flash DeepSeek-V4-Pro	enabled	enabled disabled
DeepSeek-R1-0528	enabled	enabled
DeepSeek-V3.2	disabled	enabled disabled

Tool Choices for Models

Model	Supported Tool Choice
GLM-5.1 GLM-5.2	none auto named
GLM-5	none auto
DeepSeek-V4-Flash DeepSeek-V4-Pro DeepSeek-V3.2 DeepSeek-R1-0528	none auto named
DeepSeek-V3	auto named

Example 1: Streaming Output

Example request:

Python
Curl

import requests
import json
if __name__ == '__main__':
    url = "https://api-ap-southeast-1.modelarts-maas.com/anthropic/v1/messages" # API URL
    api_key = "MAAS_API_KEY"  # Replace MAAS_API_KEY with the obtained API key.

    # Send request.
    headers = {
        'Content-Type': 'application/json',
        'Authorization': f'Bearer {api_key}'
    }
    data = {
        "model": "glm-5.2",  # Model parameter. You can change the model parameter as required.
        "stream":True,
        "messages": [
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": "Hello"}
        ]
    }
    response = requests.post(url, headers=headers, data=json.dumps(data), verify=False)
    # Print result.
    print(response.status_code)
    print(response.text)

Run the following commands in the Linux environment:

curl --request POST \
  --url https://api-ap-southeast-1.modelarts-maas.com/anthropic/v1/messages \
  --header 'content-type: application/json' \
  --header 'x-api-key: $MaaS_API_Key' \
  --data '{
  "model": "glm-5.2",
  "stream": true,
  "messages": [
    {
      "role": "system",
      "content": "You are a helpful assistant."
    },
    {
      "role": "user",
      "content": "Hello"
    }
  ]
}'

Example response:

event: message_start
data: {"type":"message_start","message":{"id":"22cadffad9724d5893e4f9ffed41a695","type":"message","role":"assistant","model":"glm-5.2","content":[],"stop_reason":null,"stop_sequence":null,"usage":{"input_tokens":0,"cache_creation_input_tokens":0,"cache_read_input_tokens":0,"output_tokens":0,"service_tier":"default"}}}

event: content_block_start
data: {"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}

event: content_block_delta
data: {"delta":{"type":"text_delta","text":"Hello!"},"type":"content_block_delta","index":0}

event: content_block_delta
data: {"delta":{"type":"text_delta","text":"Is there anything"},"type":"content_block_delta","index":0}

event: content_block_delta
data: {"delta":{"type":"text_delta","text":"I can help"},"type":"content_block_delta","index":0}

event: content_block_delta
data: {"delta":{"type":"text_delta","text":"you with"},"type":"content_block_delta","index":0}

event: content_block_delta
data: {"delta":{"type":"text_delta","text":"?"},"type":"content_block_delta","index":0}

event: content_block_stop
data: {"type":"content_block_stop","index":0}

event: message_delta
data: {"type":"message_delta","delta":{"stop_reason":"end_turn","stop_sequence":null},"usage":{"output_tokens":9,"input_tokens":18,"total_tokens":27,"cache_creation_input_tokens":0,"cache_read_input_tokens":0}}

event: message_stop
data: {"type":"message_stop"}

data: [DONE]

Example 2: Tool Calling

Example request:

Python
Curl

import requests
import json
if __name__ == '__main__':
    url = "https://api-ap-southeast-1.modelarts-maas.com/anthropic/v1/messages" # API URL
    api_key = "MAAS_API_KEY" # Replace MAAS_API_KEY with the obtained API key.

    # Send request.
    headers = {
        'Content-Type': 'application/json',
        'Authorization': f'Bearer {api_key}'
    }
    data = {
        "model": "glm-5.2",
          "thinking": {
            "type": "disabled"
          },
          "stream": False,
          "system": "You are a helpful assistant. Keep replies concise.",
          "messages": [
            {
              "role": "user",
              "content": "What is the weather in Beijing today?"
            },
            {
              "role": "assistant",
              "content": [
                {
                  "type": "tool_use",
                  "id": "call_664e61ca7977429db860d513",
                  "input": {
                    "city": "Beijing",
                    "unit": "celsius"
                  },
                  "name": "get_weather"
                }
              ]
            },
            {
              "role": "user",
              "content": [
                {
                  "type": "tool_result",
                  "tool_use_id": "call_664e61ca7977429db860d513",
                  "content": "{\"city\": \"Beijing\", \"temperature\": 26, \"unit\": \"celsius\", \"condition\": \"Sunny\"}"
                }
              ]
            }
          ],
          "tools": [
            {
              "name": "get_weather",
                "description": "Get the current weather for a specified city.",
              "input_schema": {
                "type": "object",
                "properties": {
                  "city": {
                    "type": "string",
                    "description": "City name, for example, Beijing."
                  },
                  "unit": {
                    "type": "string",
                    "enum": [
                      "celsius",
                      "fahrenheit"
                    ],
                    "description": "Temperature unit"
                  }
                },
                "required": [
                  "city"
                ]
              }
            }
          ]
    }
    response = requests.post(url, headers=headers, data=json.dumps(data), verify=False)
    # Print result.
    print(response.status_code)
    print(response.text)

Run the following commands in the Linux environment:

curl --request POST \
  --url https://api-ap-southeast-1.modelarts-maas.com/anthropic/v1/messages \
  --header 'content-type: application/json' \
  --header 'x-api-key: $MaaS_API_Key' \
  --data '{
  "model": "glm-5.2",
  "thinking": {
    "type": "disabled"
  },
  "stream": false,
  "system": "You are a helpful assistant. Keep replies concise.",
  "messages": [
    {
      "role": "user",
      "content": "What is the weather in Beijing today?"
    },
    {
      "role": "assistant",
      "content": [
        {
          "type": "tool_use",
          "id": "call_664e61ca7977429db860d513",
          "input": {
            "city": "Beijing",
            "unit": "celsius"
          },
          "name": "get_weather"
        }
      ]
    },
    {
      "role": "user",
      "content": [
        {
          "type": "tool_result",
          "tool_use_id": "call_664e61ca7977429db860d513",
            "content": "{\"city\": \"Beijing\", \"temperature\": 26, \"unit\": \"celsius\", \"condition\": \"Sunny\"}"
        }
      ]
    }
  ],
  "tools": [
    {
      "name": "get_weather",
       "description": "Get the current weather for a specified city.",
      "input_schema": {
        "type": "object",
        "properties": {
          "city": {
            "type": "string",
              "description": "City name, for example, Beijing."
          },
          "unit": {
            "type": "string",
            "enum": [
              "celsius",
              "fahrenheit"
            ],
              "description": "Temperature unit"
          }
        },
        "required": [
          "city"
        ]
      }
    }
  ]
}'

Example response:

{
  "content": [
    {
      "text": "Today's weather in Beijing: **Sunny**, temperature **26°C**. The weather is great for travel! ☀️",
      "type": "text"
    }
  ],
  "id": "1d3756ed5704491f980c2fd40054bfd7",
  "model": "glm-5.2",
  "role": "assistant",
  "stop_reason": "end_turn",
  "stop_sequence": null,
  "type": "message",
  "usage": {
    "cache_creation_input_tokens": 0,
    "cache_read_input_tokens": 0,
    "input_tokens": 250,
    "output_tokens": 23,
    "service_tier": "default",
    "total_tokens": 273
  }
}

Parent Topic: Sending a Chat Request (Chat/Post)

Previous topic: OpenAI-compatible APIs

Next topic: Obtaining the Model List (Models/GET)

Feedback

Was this page helpful?

Helpful Not helpful

Provide feedback

Thank you very much for your feedback. We will continue working to improve the documentation.See the reply and handling status in My Cloud VOC.

The system is busy. Please try again later.

For any further questions, feel free to contact us through the chatbot.

Chatbot