GLM-5.3 Flash

GLM-5.3 Flash uses HopBase's OpenAI Chat Completions compatible protocol. Use https://api.hop-base.com/v1 and set the model to glm-5.3-flash.

Model and specifications

Endpoint and Base URL

Group selection and billing

• glm-5.3-flash is sold through the GLM 5.3 Flash 专线 (GLM 5.3 Flash dedicated line) group. When creating an API key under API Keys in the console, set the key's group to this one; only a key bound to it can call the model.
• One key is bound to one group. Keys bound to other groups get a clear 404 when calling glm-5.3-flash — create a new key on this group, or rebind an existing key in the console.
• Rates take Zhipu's official list price for this model as their basis and are published on the pricing page; your own effective rate is shown in the signed-in model catalog.
• glm-5.3 keeps its own groups and is unaffected; see GLM-5.3.

Recommended parameters

Zhipu publishes these recommended values for this model:

curl

curl https://api.hop-base.com/v1/chat/completions \
  -H "Authorization: Bearer sk-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.3-flash",
    "temperature": 1,
    "top_p": 0.95,
    "reasoning_effort": "max",
    "messages": [{"role": "user", "content": "Hello, introduce yourself"}]
  }'

Streaming and usage

Send "stream": true. To receive token counts, add stream_options.include_usage; the final chunk then carries a usage object. When the request also carries tools, set tool_stream: true so tool-call arguments stream incrementally instead of arriving in one block.

curl https://api.hop-base.com/v1/chat/completions \
  -H "Authorization: Bearer sk-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.3-flash",
    "stream": true,
    "stream_options": {"include_usage": true},
    "tool_stream": true,
    "messages": [{"role": "user", "content": "Write a haiku about latency"}]
  }'

Tool calling

Function calling works over the standard OpenAI fields. Pass tools, read choices[].message.tool_calls from the reply, run the function, then send the result back as a tool message.

curl https://api.hop-base.com/v1/chat/completions \
  -H "Authorization: Bearer sk-your-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "glm-5.3-flash",
    "messages": [{"role": "user", "content": "What is the weather in Tokyo?"}],
    "tools": [{
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Get the current weather for a city",
        "parameters": {
          "type": "object",
          "properties": {"city": {"type": "string"}},
          "required": ["city"]
        }
      }
    }]
  }'

Thinking output

The model always thinks; there is no switch to turn it off. The thinking trace is returned in reasoning_content alongside the answer in content, and in streaming mode it arrives as delta.reasoning_content. Render it separately, or drop it, but do not concatenate it into the answer. Size max_tokens for the answer and the trace together.

OpenAI SDK / Python

from openai import OpenAI

client = OpenAI(
    base_url="https://api.hop-base.com/v1",
    api_key="sk-your-key",
)

resp = client.chat.completions.create(
    model="glm-5.3-flash",
    temperature=1,
    top_p=0.95,
    reasoning_effort="max",
    messages=[{"role": "user", "content": "Explain where GLM-5.3 Flash works best"}],
)
print(resp.choices[0].message.content)
print(resp.usage)

FAQ

• Calling glm-5.3-flash returns 404 model not found. The key is not on the GLM 5.3 Flash dedicated line group; create a new key on that group or rebind the existing one in the console.
• Streaming returns no usage. Add "stream_options": {"include_usage": true}; without it the final chunk carries no token counts.
• Thinking cannot be disabled. This model always thinks. Use reasoning_effort to pick the depth instead of trying to switch thinking off.