Create Chat Completions

Generates one or more responses from a language model using the conversation context provided by the user. This core endpoint supports question answering, conversation, content creation, and both streaming and non-streaming responses.

Authorization

bearer

AuthorizationBearer <token>

In: header

Header Parameters

Content-Type*string

Accept*string

Authorization?string

Request Body

application/json

TypeScript Definitions

Use the request body type in TypeScript.

model*string

The ID of the model to use. See the model endpoint compatibility table for details about which models support the Chat API.

messages*array<>

A list of the messages in the conversation so far.

temperature?integer

The sampling temperature to use, from 0 to 2. Higher values such as 0.8 make the output more random, while lower values such as 0.2 make it more focused and deterministic. We generally recommend changing either temperature or top_p, but not both.

top_p?integer

An alternative to temperature sampling called nucleus sampling, in which the model considers tokens whose cumulative probability mass is top_p. For example, 0.1 means that only the tokens comprising the top 10% of the probability mass are considered. We generally recommend changing either top_p or temperature, but not both.

n?integer

Defaults to 1. The number of chat completion choices to generate for each input message.

stream?boolean

Defaults to false. When enabled, partial message deltas are sent as data-only server-sent events as they become available, and a data: [DONE] message terminates the stream.

stop?string

Defaults to null. Up to four sequences where the API will stop generating further tokens.

max_tokens?integer

Defaults to infinity. The maximum number of tokens to generate in the chat completion.

The combined length of the input and generated tokens is limited by the model's context window.

presence_penalty?number

A number from -2.0 to 2.0. Positive values penalize new tokens based on whether they have appeared in the text so far, increasing the model's likelihood of discussing new topics. Learn more about frequency and presence penalties.

frequency_penalty?number

Defaults to 0. A number from -2.0 to 2.0. Positive values penalize new tokens based on their frequency in the text so far, reducing the model's likelihood of repeating the same lines.

logit_bias?null

Modifies the likelihood that specified tokens will appear in the completion.

Accepts a JSON object that maps token IDs, as defined by the tokenizer, to bias values from -100 to 100. The bias is mathematically added to the logits generated by the model before sampling. The exact effect varies by model, but values from -1 to 1 should decrease or increase the likelihood of selecting the associated token, while values such as -100 or 100 should effectively prohibit or exclusively select it.

user?string

A unique identifier representing your end user, which can help OpenAI monitor and detect abuse. Learn more.

response_format?

An object specifying the format that the model must output. Set { "type": "json_object" } to enable JSON mode, which ensures that the model generates valid JSON. Important: When using JSON mode, you must also instruct the model to produce JSON in a system or user message. Otherwise, the model may generate an unending stream of whitespace until it reaches the token limit, causing increased latency and an apparently stalled request. The message content may also be partially cut off when finish_reason="length", which indicates that generation exceeded max_tokens or the conversation exceeded the maximum context length.

seen?integer

This feature is in beta. When specified, the system makes a best effort to sample deterministically so that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed; use the system_fingerprint response parameter to monitor backend changes.

tools*array<string>

A list of tools the model may call. Currently, only functions are supported as tools. Use this field to provide functions for which the model may generate JSON input.

tool_choice*

Controls which function, if any, the model calls. none means the model will generate a message without calling a function. auto lets the model choose between generating a message and calling a function. Use {"type": "function", "function": {"name": "my_function"}} to force the model to call that function. The default is none when no functions are present and auto when functions are present.

Response Body

`application/json`

`text/event-stream`

curl -X POST "https://example.com/v1/chat/completions" \  -H "Content-Type: application/json" \  -H "Accept: application/json" \  -H "Authorization: Bearer {{YOUR_API_KEY}}" \  -H "Content-Type: application/json" \  -d '{    "model": "Qwen3-30B-A3B-Instruct-2507",    "messages": [      {        "role": "system",        "content": "You are a helpful assistant."      },      {        "role": "user",        "content": "Hello!"      }    ],    "top_k": 0,    "min_p": 0,    "stream": true  }'

{  "id": "chatcmpl-123",  "object": "chat.completion",  "created": 1677652288,  "choices": [    {      "index": 0,      "message": {        "role": "assistant",        "content": "\n\nHello there, how may I assist you today?"      },      "finish_reason": "stop"    }  ],  "usage": {    "prompt_tokens": 9,    "completion_tokens": 12,    "total_tokens": 21  }}