Function Calling with Chinese AI Models

title: "Function Calling with Chinese AI Models: DeepSeek, GLM & Kimi"

published: true

tags: function-calling, tool-use, deepseek, glm, kimi

canonical_url: https://aiwave.live/blog/function-calling-chinese-ai-models

description: "Function calling support across DeepSeek V4, GLM-5, and Kimi K3. Format comparison and working Python examples for each model."

Function calling (also called tool use) lets LLMs execute structured actions: query databases, call APIs, trigger workflows. All three major Chinese model providers support it through OpenAI-compatible APIs.

Available through AIWave with model names deepseek-chat, glm-4-plus, kimi-k3.

Format

All three accept the same tools parameter format as OpenAI:

{
  "model": "deepseek-chat",
  "messages": [...],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Get current weather for a city",
        "parameters": {
          "type": "object",
          "properties": {
            "city": {"type": "string"},
            "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}
          },
          "required": ["city"]
        }
      }
    }
  ]
}

The model returns a tool_calls array with the function name and arguments. Your code executes the function, then sends the result back.

Implementation

from openai import OpenAI
import json

client = OpenAI(api_key="***", base_url="https://aiwave.live/v1")

def get_weather(city, unit="celsius"):
    # Your actual API call here
    return {"temp": 22, "condition": "sunny"}

tools = [{
    "type": "function",
    "function": {
        "name": "get_weather",
        "description": "Get weather for a city",
        "parameters": {
            "type": "object",
            "properties": {
                "city": {"type": "string"},
                "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]}
            },
            "required": ["city"]
        }
    }
}
}]

messages = [{"role": "user", "content": "What's the weather in Tokyo?"}]

response = client.chat.completions.create(
    model="deepseek-chat",
    messages=messages,
    tools=tools
)

# Extract and execute function call
tool_call = response.choices[0].message.tool_calls[0]
args = json.loads(tool_call.function.arguments)
result = get_weather(args["city"])

# Send result back
messages.append(response.choices[0].message)
messages.append({"role": "tool", "tool_call_id": tool_call.id, "content": json.dumps(result)})

final = client.chat.completions.create(
    model="deepseek-chat",
    messages=messages
)
print(final.choices[0].message.content)

Model Differences

All three work with the same code above. Switch models by changing the model parameter.

Real-World Use Cases

Function calling is available on all AIWave models. The $0.20 starter credit covers hundreds of tool-calling turns to prototype your agent.