openai.createChatCompletion

Returns a predicted chat completion for the given messages.

Syntax

createChatCompletion(params)

Parameters

• params

  An object with request parameters:

  • model optional

    An identifier of the model to use (Defaults to "gpt-3.5-turbo").

  • messages

    A an array of message objects with the following properties:

    • role

    A string describing the role ("user", "assistant", "system", or "function").

    • content

    A string that contains the message text.

    • name optional

    The message author name, or the name of the function that generated content.

    • function_call optional

    The name and the arguments of a function call.

  • functions optional

    An array of function description objects that the model can use to generate JSON inputs for. A function object contains:

    • name

    The function name.

    • description optional

    A description of the function.

    • parameters optional

    The function parameters described using JSON Schema.

  • function_call optional

    The way of responding to function calls: a string ("none", "auto") or an object containing a single key "name" and the name of the function to call.

  • max_tokens optional

    A maximum number of tokens to generate (defaults to 16).

  • temperature optional

    A temperature from 0 to 2 (defaults to 1). Higher values give more random results.

  • top_p optional

    An alternative to temperature.

  • n optional

    The number of completions to generate.

  • logprobs optional

    The number of most likely tokens for which to include the log probabilities (defaults to 5, maximum is 5).

  • echo optional

    Include the prompt in the response.

  • stop optional

    An array of sentences or a string where the API will stop generating tokens (maximum is 4).

  • presence_penalty optional

    A number between -2.0 and 2.0 (defaults to 0).

  • frequency_penalty optional

    A number between -2.0 and 2.0 (defaults to 0).

  • logit_bias optional

    The likelihood of specified tokens appearing in the completion.

  • user optional

    A unique identifier of the end user.

Return value

An object containing response from OpenAI API.

Example

await openai.createChatCompletion({
  model: "gpt-3.5-turbo",
  messages: [
    {role: "user", content: "Hello!"},
    {role: "assistant", content: "Hello! How can I assist you today?"},
    {role: "user", content: "What's the weight of Dacia Sandero?"}
  ],
})

{
  id: chatcmpl-6pOKSz35WXoPaQyhsb8b4FFqLqe3G,
  object: chat.completion,
  created: 1677706268,
  model: gpt-3.5-turbo-0301,
  usage: {
           prompt_tokens: 38,
           completion_tokens: 33,
           total_tokens: 71
         },
  choices: [{
             message: {
                        role: assistant,
                        content: The weight of Dacia Sandero varies depending on the specific model and equipment, but generally it ranges from 945 kg to 1,180 kg.
                      },
             finish_reason: stop,
             index: 0
           }]
}

External resources

• OpenAI reference