Skip to main content
Different model families have different defaults and constraints for Chat Completions API parameters. For the full model list, see the Model List.

​
Parameter Comparison

Parameterkimi-k2.5kimi-k2 serieskimi-k2-thinking seriesmoonshot-v1 series
temperatureCannot be modified0.61.00.0
top_p0.95 Cannot be modified1.01.01.0
n1 Cannot be modified1 (max 5)1 (max 5)1 (max 5)
presence_penalty0 Cannot be modified0 (modifiable)0 (modifiable)0 (modifiable)
frequency_penalty0 Cannot be modified0 (modifiable)0 (modifiable)0 (modifiable)
thinkingSupported———
When temperature is close to 0, n can only be 1. Otherwise, the API returns invalid_request_error.

​
kimi-k2.5 — thinking Parameter

kimi-k2.5 supports the thinking parameter to control whether deep thinking is enabled. Accepts {"type": "enabled"} or {"type": "disabled"}. Since the OpenAI SDK doesn’t have a native thinking parameter, use extra_body: 
completion = client.chat.completions.create(
    model="kimi-k2.5",
    messages=[
        {"role": "user", "content": "Hello"}
    ],
    extra_body={
        "thinking": {"type": "disabled"}
    },
    max_tokens=1024*32,
)