GLM 5 API

OpenAI-compatible chat completions, models, API keys, billing, streaming, and function calling.

The GLM 5 API is an OpenAI-compatible developer API for chat completions. Use your GLM 5 API key with the /api/v1 base URL, choose a public model ID, and pay with the same credits used by Chat.

Base URL

https://glm5.app/api/v1

Authentication

Create an API key in /settings/apikeys. Plaintext keys are shown once and stored as hashes.

Send the key as a bearer token:

Authorization: Bearer sk-glm5-...

Quickstart

curl https://glm5.app/api/v1/chat/completions \
  -H "Authorization: Bearer $GLM5_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.2",
    "messages": [
      {
        "role": "user",
        "content": "Write a short launch checklist for a developer API."
      }
    ]
  }'

from openai import OpenAI

client = OpenAI(
    api_key="sk-glm5-...",
    base_url="https://glm5.app/api/v1",
)

completion = client.chat.completions.create(
    model="glm-5.2",
    messages=[
        {"role": "user", "content": "Write a short launch checklist."}
    ],
)

print(completion.choices[0].message.content)

import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.GLM5_API_KEY,
  baseURL: "https://glm5.app/api/v1",
});

const completion = await client.chat.completions.create({
  model: "glm-5.2",
  messages: [
    { role: "user", content: "Write a short launch checklist." },
  ],
});

console.log(completion.choices[0]?.message?.content);

Chat Completions

POST /chat/completions

Parameters

Parameter	Type	Required	Default	Notes
`model`	string	Yes	-	One of `glm-5.2`, `glm-5`, `kimi-k2`, `deepseek-r1`.
`messages`	array	Yes	-	OpenAI chat messages: `system`, `user`, `assistant`, and `tool`.
`max_tokens`	integer	No	`8192`	Capped by the selected model.
`temperature`	number	No	provider default	Sampling temperature.
`top_p`	number	No	provider default	Nucleus sampling.
`top_k`	integer	No	provider default	Restrict choices to the top K tokens.
`seed`	integer	No	-	Deterministic sampling when supported.
`stop`	string or array	No	-	Stop sequences.
`stream`	boolean	No	`false`	Streams OpenAI-style SSE chunks.
`tools`	array	No	-	Function tools. Supported by `glm-5.2`, `glm-5`, and `kimi-k2`.
`tool_choice`	string or object	No	`auto`	`auto`, `none`, `required`, or a named function choice.

Streaming

Set stream to true to receive Server-Sent Events:

curl https://glm5.app/api/v1/chat/completions \
  -H "Authorization: Bearer $GLM5_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.2",
    "stream": true,
    "messages": [{"role": "user", "content": "Say hello in one sentence."}]
  }'

The stream emits OpenAI-compatible chat.completion.chunk events and ends with:

data: [DONE]

Function Calling

Function calling works with multi-turn tool loops. Send tools, receive tool_calls, execute the tool in your application, then send the tool result back with role: "tool".

{
  "model": "glm-5.2",
  "messages": [
    {
      "role": "user",
      "content": "What is the weather in Shanghai?"
    }
  ],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Get current weather for a city.",
        "parameters": {
          "type": "object",
          "properties": {
            "city": { "type": "string" }
          },
          "required": ["city"]
        }
      }
    }
  ]
}

Send the result back:

{
  "model": "glm-5.2",
  "messages": [
    {
      "role": "user",
      "content": "What is the weather in Shanghai?"
    },
    {
      "role": "assistant",
      "tool_calls": [
        {
          "id": "call_123",
          "type": "function",
          "function": {
            "name": "get_weather",
            "arguments": "{\"city\":\"Shanghai\"}"
          }
        }
      ]
    },
    {
      "role": "tool",
      "tool_call_id": "call_123",
      "content": "Shanghai is 24C and clear."
    }
  ]
}

Models

GET /models

Returns the public model IDs available through GLM 5.

Model	Tools	Max output tokens	Input price	Output price
`glm-5.2`	Yes	8192	`$2.50 / 1M tokens`	`$7.50 / 1M tokens`
`glm-5`	Yes	8192	`$1.50 / 1M tokens`	`$5.00 / 1M tokens`
`kimi-k2`	Yes	8192	`$1.50 / 1M tokens`	`$6.00 / 1M tokens`
`deepseek-r1`	No	8192	`$1.75 / 1M tokens`	`$6.50 / 1M tokens`

Billing

API usage is billed from your existing GLM 5 credits. Credits power both Chat and API calls; there is no separate API balance or separate API subscription tier.

Credits are reserved before an upstream request is made, then reconciled after final usage is known. Failed requests are refunded.

Rate Limits

The default per-key limit is 60 requests per minute. If a key exceeds its limit, the API returns 429 rate_limit_exceeded.

Errors

Errors use the OpenAI shape:

{
  "error": {
    "message": "Invalid API key provided.",
    "type": "invalid_request_error",
    "code": "invalid_api_key",
    "param": null
  }
}

Status	Code	Meaning
400	`invalid_request_error`	Malformed request body or messages.
400	`unsupported_parameter`	A parameter is unsupported for the selected model.
401	`invalid_api_key`	Missing or invalid API key.
402	`insufficient_quota`	Not enough credits.
404	`model_not_found`	Unknown model ID.
429	`rate_limit_exceeded`	Per-key rate limit exceeded.
503	`service_unavailable`	Public API is temporarily disabled.
500	`internal_error`	Unexpected server error.

GLM 5 API

On this page