curl --request POST \
  --url https://direct.evolink.ai/v1/messages \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "glm-5.3",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Hello, world"
    }
  ]
}
'

{
  "id": "msg_0842a705-9d0b-4eaa-b12d-09a4106326c5",
  "type": "message",
  "role": "assistant",
  "model": "glm-5.3",
  "content": [
    {
      "type": "thinking",
      "thinking": "The user asked to greet them with one word, so answering \"Hi\" will do.",
      "signature": ""
    },
    {
      "type": "text",
      "text": "Hi."
    }
  ],
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 18,
    "output_tokens": 101,
    "cache_creation_input_tokens": 0,
    "cache_read_input_tokens": 0,
    "prompt_tokens_details": {
      "cached_tokens": 0
    }
  }
}

Messages API

GLM All-Model API - Messages Reference

Call GLM series models through the Anthropic Messages protocol, choosing the specific model with the model parameter
The request and response structures match the Anthropic API
System prompt: passed through the top-level system field
Thinking mode: thinking is on by default across the series and is returned in content[type=thinking] blocks; only glm-5.2 can turn it off with thinking.type=disabled
Streaming: SSE event stream
Tool calling: compatible with the Anthropic tool_use / tool_result flow
Image input: supported only by glm-5.3-flash and glm-5.3-flashx; see the messages field for details

POST

messages

curl --request POST \
  --url https://direct.evolink.ai/v1/messages \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "model": "glm-5.3",
  "max_tokens": 1024,
  "messages": [
    {
      "role": "user",
      "content": "Hello, world"
    }
  ]
}
'

{
  "id": "msg_0842a705-9d0b-4eaa-b12d-09a4106326c5",
  "type": "message",
  "role": "assistant",
  "model": "glm-5.3",
  "content": [
    {
      "type": "thinking",
      "thinking": "The user asked to greet them with one word, so answering \"Hi\" will do.",
      "signature": ""
    },
    {
      "type": "text",
      "text": "Hi."
    }
  ],
  "stop_reason": "end_turn",
  "usage": {
    "input_tokens": 18,
    "output_tokens": 101,
    "cache_creation_input_tokens": 0,
    "cache_read_input_tokens": 0,
    "prompt_tokens_details": {
      "cached_tokens": 0
    }
  }
}

BaseURL: The default BaseURL is https://direct.evolink.ai, which has better support for text models and long-lived connections. https://api.evolink.ai is the primary endpoint for multimodal services and serves as a fallback address for text models.

Thinking settings. Thinking is enabled by default for all models. Response thinking blocks are billed as output tokens; signature may be an empty string.

glm-5.2: thinking.type=disabled turns thinking off.
glm-5.3 / glm-5.3-flash / glm-5.3-flashx: thinking is always enabled. Passing disabled retains the default thinking mode and its output-token charges.

thinking.budget_tokens and thinking.effort do not set GLM thinking levels. See the top-level reasoning_effort field for compatible values. A setting that disables thinking on glm-5.2 does not disable it after switching to the 5.3 series.

reasoning_effort: GLM reasoning-level compatibility field at the request top level, outside thinking. The endpoint accepts this field but does not guarantee that it controls actual thinking effort. Compatibility rules for glm-5.3 / glm-5.3-flash / glm-5.3-flashx

Supplied value	Compatible value
`low` / `high` / `max`	Unchanged
`xhigh`	`max`
`medium`	`high`
`minimal` / `none`	`low`; thinking remains enabled

minimal and none do not disable thinking on the 5.3 series. Thinking tokens are billed as output. Unrecognized values are unchanged and have no compatibility mapping; use the listed values. These mappings do not apply to glm-5.2. To control thinking effort, use reasoning_effort in Chat Completions or reasoning.effort in Responses.

Only glm-5.3-flash and glm-5.3-flashx support image input. Image blocks sent to glm-5.3 or glm-5.2 may return 200 even though the model cannot read them. Choose an image-capable model for image understanding.

Authorizations

Authorization

string

header

required

##All interfaces require authentication using a Bearer Token##

Get an API Key:

Visit the API Key management page to obtain your API Key

Add it to the request header when using:

Authorization: Bearer YOUR_API_KEY

Note: EvoLink uses Bearer Token authentication uniformly for /v1/messages.

Body

application/json

model

enum<string>

default:glm-5.3

required

Model to call:

Model ID	Positioning	Thinking can be turned off	Image input
`glm-5.3`	Flagship model, with across-the-board gains on complex software engineering and agent tasks; 1M context	Cannot be turned off	Not supported
`glm-5.3-flash`	Lightweight multimodal model, extremely low cost with native vision; 1M context	Cannot be turned off	Supported
`glm-5.3-flashx`	Multimodal model with image input; 1M context	Cannot be turned off	Supported
`glm-5.2`	Previous-generation flagship, complex reasoning and very long context; 1M context	Can be turned off (`thinking.type=disabled`)	Not supported

Available options:

glm-5.3,

glm-5.3-flash,

glm-5.3-flashx,

glm-5.2

Example:

"glm-5.3"

messages

object[]

required

Conversation messages, alternating between user and assistant. Include at least one message; the last message is usually from the user. Include earlier messages for multi-turn context.

Image input: only glm-5.3-flash and glm-5.3-flashx support image blocks in the content array. Use text only with glm-5.3 and glm-5.2. An HTTP 200 response does not mean that an unsupported model read the image.

Minimum array length: 1

Show child attributes

max_tokens

integer

Upper limit on the length of the generated content (in tokens)

Note:

The GLM series supports up to 131,072 tokens (128K) of output; setting at least 1024 is recommended
Tokens produced by thinking also count towards this limit
When the limit is reached the content is truncated and the response carries stop_reason=max_tokens

Required range: 1 <= x <= 131072

Example:

1024

system

System prompt, used to set the AI's role and behavior

Notes:

Supports a string or an array of content blocks
Passed via the top-level system field (do not place it inside messages)
The model follows the system constraints
An overly long system may be truncated: For long context, place it in messages rather than piling everything into system

Example:

"You are a helpful assistant."

temperature

number

Sampling temperature

Notes:

Higher values make output more varied, lower values more deterministic
Recommended range [0, 1]

Required range: 0 <= x <= 1

Example:

1

top_p

number

Nucleus sampling threshold

Notes:

Range [0, 1]
It is recommended not to adjust temperature and top_p at the same time

Required range: 0 <= x <= 1

Example:

0.9

top_k

integer

Sample only from the K highest-probability tokens (an Anthropic-specific parameter)

Notes:

Smaller values make output more deterministic, larger values make candidates more diverse

Required range: x >= 0

Example:

10

stop_sequences

string[]

Custom stop sequences: generation stops when it hits any of these strings

Notes:

Hitting one truncates output, and content before the hit is returned normally
Note: When a stop sequence is hit, the GLM series returns stop_reason as end_turn (rather than the Anthropic-standard stop_sequence), and the response does not include a stop_sequence field. If your client relies on stop_reason=="stop_sequence" to detect a hit, special handling is required

Example:

["\n\n"]

stream

boolean

default:false

Whether to return via SSE streaming

true: Server-Sent Events streaming (standard Anthropic event sequence: message_start / content_block_start / content_block_delta / message_delta / message_stop)
false: Returns the complete response all at once (default)

Example:

false

thinking

object

Thinking settings. Thinking is enabled by default for all models. Response thinking blocks are billed as output tokens; signature may be an empty string.

glm-5.2: thinking.type=disabled turns thinking off.
glm-5.3 / glm-5.3-flash / glm-5.3-flashx: thinking is always enabled. Passing disabled retains the default thinking mode and its output-token charges.

thinking.budget_tokens and thinking.effort do not set GLM thinking levels. See the top-level reasoning_effort field for compatible values. A setting that disables thinking on glm-5.2 does not disable it after switching to the 5.3 series.

Show child attributes

reasoning_effort

enum<string>

GLM reasoning-level compatibility field at the request top level, outside thinking. The endpoint accepts this field but does not guarantee that it controls actual thinking effort.

Compatibility rules for glm-5.3 / glm-5.3-flash / glm-5.3-flashx

Supplied value	Compatible value
`low` / `high` / `max`	Unchanged
`xhigh`	`max`
`medium`	`high`
`minimal` / `none`	low; thinking remains enabled

minimal and none do not disable thinking on the 5.3 series. Thinking tokens are billed as output. Unrecognized values are unchanged and have no compatibility mapping; use the listed values. These mappings do not apply to glm-5.2.

To control thinking effort, use reasoning_effort in Chat Completions or reasoning.effort in Responses.

Available options:

max,

xhigh,

high,

medium,

low,

minimal,

none

tools

object[]

The list of tool definitions

Notes:

Follows the Anthropic tool definition spec
input_schema uses a JSON Schema object
The model returns standard tool_use blocks with stop_reason=tool_use

Show child attributes

tool_choice

object

Tool selection strategy

Show child attributes

metadata

object

Request metadata

Show child attributes

Response

Message object

Anthropic-style message response

string

The message's unique ID (format: msg_<uuid>)

type

enum<string>

Response object type

Available options:

message

role

enum<string>

Available options:

assistant

model

string

Model actually used

Example:

"glm-5.3"

content

object[]

The list of response content blocks

Possible block types:

thinking: the reasoning process (when thinking is enabled, which is the default)
text: the final answer text
tool_use: a tool call initiated by the model

Show child attributes

stop_reason

enum<string>

Stop reason

end_turn: natural completion (also returned when stop_sequences is hit)
max_tokens: reached the max_tokens limit
tool_use: the model triggered a tool call

Available options:

end_turn,

max_tokens,

tool_use

usage

object

Token usage statistics (Anthropic specification)

Show child attributes

Image Series

Video Series

Audio Series

Text Series

Account Management

Task Management

File Management

GLM All-Model API - Messages Reference

Authorizations

Body

Response