API Reference

Request response from the model

post
https://integrate.api.nvidia.com/v1/chat/completions

Invokes inference using the model chat parameters.

Recent Requests
Log in to see full request history
TimeStatusUser Agent
Retrieving recent requests…
LoadingLoading…
Body Params
messages
array of objects
required
length ≥ 1

A list of messages comprising the conversation so far.

Messages*
string
enum
required

The role of the message's author.

Allowed:
string | array | null
required

The contents of the message.

To pass media (only with role=user):

- Use content parts. Text is passed with type=text.

- Images are passed with type=image_url; set image_url.url to an image URL or a base64 data URI like data:image/{format};base64,{base64encodedimage}.

For system and assistant roles, the object list format is not supported.

Optional participant or tool name.

Tool call ID that this tool message is responding to.

array | null

Tool calls generated by an assistant message.

Optional preserved reasoning content from a previous assistant turn.

string
required
Defaults to moonshotai/kimi-k3

The model to use.

array | null

Optional OpenAI-compatible tool definitions for function calling.

1 to 65536

The maximum number of tokens that can be generated.

-9007199254740991 to 9007199254740991

Changing the seed will produce a different response with similar characteristics. Fixing the seed will reproduce the same results if all other parameters are also kept constant.

If set, partial message deltas will be sent, like in ChatGPT. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: [DONE]

object | null

Optional OpenAI-compatible stream options, such as include_usage.

0 to 1

What sampling temperature to use, between 0 and 1. Higher values make the output more random; lower values make it more focused and deterministic. Recommended for Kimi-K3: 1.0. Note: the other sampling parameters (top_p, presence_penalty, frequency_penalty, n) are fixed by the model and are not exposed.

enum
Defaults to max

Controls how long the model reasons before answering: low, high, or max. Per-request override of the serve-time default; unset falls back to the model default (max).

Allowed:
Headers
string
enum
Defaults to application/json

Generated from available response content types

Allowed:
Responses

Updated about 1 hour ago

Did this page help you?
Language
Credentials
Bearer
LoadingLoading…
Response
Click Try It! to start a request and see the response here! Or choose an example:
application/json
text/event-stream

Updated about 1 hour ago

Did this page help you?
country_code