kimi-k2-preview
Model Overview
A mixture-of-experts model with strong reasoning, coding, and agentic capabilities.
How to Make a Call
API Schema
Creates a chat completion using a language model, allowing interactive conversation by predicting the next response based on the given chat history. This is useful for AI-driven dialogue systems and virtual assistants.
An upper bound for the number of tokens that can be generated for a completion, including visible output tokens and reasoning tokens.
512The maximum number of tokens that can be generated in the chat completion. This value can be used to control costs for text generated via API.
512If set to True, the model response data will be streamed to the client as it is generated using server-sent events.
falseControls which (if any) tool is called by the model. none means the model will not call any tool and instead generates a message. auto means the model can pick between generating a message or calling one or more tools. required means the model must call one or more tools. Specifying a particular tool via {"type": "function", "function": {"name": "my_function"}} forces the model to call that tool. none is the default when no tools are present. auto is the default if tools are present.
none means the model will not call any tool and instead generates a message. auto means the model can pick between generating a message or calling one or more tools. required means the model must call one or more tools.
Whether to enable parallel function calling during tool use.
How many chat completion choices to generate for each input message. Note that you will be charged based on the number of generated tokens across all of the choices. Keep n as 1 to minimize costs.
What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. We generally recommend altering this or top_p but not both.
An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. We generally recommend altering this or temperature but not both.
Up to 4 sequences where the API will stop generating further tokens. The returned text will not contain the stop sequence.
Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.
Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics.
An object specifying the format that the model must output.
POST /v1/chat/completions HTTP/1.1
Host: api.aimlapi.com
Authorization: Bearer <YOUR_AIMLAPI_KEY>
Content-Type: application/json
Accept: */*
Content-Length: 490
{
"model": "moonshot/kimi-k2-preview",
"messages": [
{
"role": "user",
"content": "text",
"name": "text"
}
],
"max_completion_tokens": 512,
"max_tokens": 512,
"stream": false,
"stream_options": {
"include_usage": true
},
"tools": [
{
"type": "function",
"function": {
"description": "text",
"name": "$web_search",
"parameters": null,
"required": [
"text"
]
}
}
],
"tool_choice": "none",
"parallel_tool_calls": true,
"n": 1,
"temperature": 1,
"top_p": 1,
"stop": "text",
"frequency_penalty": 1,
"presence_penalty": 1,
"response_format": {
"type": "text"
}
}No content
Code Example #1: Chat Completion
import requests
import json # for getting a structured output with indentation
response = requests.post(
"https://api.aimlapi.com/v1/chat/completions",
headers={
"Content-Type":"application/json",
# Insert your AIML API Key instead of <YOUR_AIMLAPI_KEY>:
"Authorization":"Bearer <YOUR_AIMLAPI_KEY>",
"Content-Type":"application/json"
},
json={
"model":"moonshot/kimi-k2-preview",
"messages":[
{
"role":"user",
# Insert your question for the model here, instead of Hello:
"content":"Hello!"
}
]
}
)
data = response.json()
print(json.dumps(data, indent=2, ensure_ascii=False))Code Example #2: Web Search
import json
import requests
from typing import Dict, Any
# Insert your AIML API Key instead of <YOUR_AIMLAPI_KEY>:
API_KEY = "<YOUR_AIMLAPI_KEY>"
BASE_URL = "https://api.aimlapi.com/v1"
HEADERS = {
"Authorization": f"Bearer {API_KEY}",
"Content-Type": "application/json",
}
def search_impl(arguments: Dict[str, Any]) -> Any:
return arguments
def chat(messages):
url = f"{BASE_URL}/chat/completions"
payload = {
"model": "moonshot/kimi-k2-preview",
"messages": messages,
"temperature": 0.6,
"tools": [
{
"type": "builtin_function",
"function": {"name": "$web_search"},
}
]
}
response = requests.post(url, headers=HEADERS, json=payload)
response.raise_for_status()
return response.json()["choices"][0]
def main():
messages = [
{"role": "system", "content": "You are Kimi."},
{"role": "user", "content": "Please search for Moonshot AI Context Caching technology and tell me what it is in English."}
]
finish_reason = None
while finish_reason is None or finish_reason == "tool_calls":
choice = chat(messages)
finish_reason = choice["finish_reason"]
message = choice["message"]
if finish_reason == "tool_calls":
messages.append(message)
for tool_call in message["tool_calls"]:
tool_call_name = tool_call["function"]["name"]
tool_call_arguments = json.loads(tool_call["function"]["arguments"])
if tool_call_name == "$web_search":
tool_result = search_impl(tool_call_arguments)
else:
tool_result = f"Error: unable to find tool by name '{tool_call_name}'"
messages.append({
"role": "tool",
"tool_call_id": tool_call["id"],
"name": tool_call_name,
"content": json.dumps(tool_result),
})
print(message["content"])
if __name__ == "__main__":
main()Last updated
Was this helpful?