Call the OpenAI API to interact with ChatGPT or o-reasoning models

Usage

chatgpt(
  .llm,
  .model = "gpt-4o",
  .max_tokens = 1024,
  .temperature = NULL,
  .top_p = NULL,
  .top_k = NULL,
  .frequency_penalty = NULL,
  .presence_penalty = NULL,
  .api_url = "https://api.openai.com/",
  .timeout = 60,
  .verbose = FALSE,
  .wait = TRUE,
  .json = FALSE,
  .min_tokens_reset = 0L,
  .stream = FALSE,
  .dry_run = FALSE
)

Arguments

.llm: An existing LLMMessage object or an initial text prompt.
.model: The model identifier (default: "gpt-4o").
.max_tokens: The maximum number of tokens to generate (default: 1024).
.temperature: Control for randomness in response generation (optional).
.top_p: Nucleus sampling parameter (optional).
.top_k: Top k sampling parameter (optional).
.frequency_penalty: Controls repetition frequency (optional).
.presence_penalty: Controls how much to penalize repeating content (optional)
.api_url: Base URL for the API (default: https://api.openai.com/v1/completions).
.timeout: Request timeout in seconds (default: 60).
.verbose: Should additional information be shown after the API call
.wait: Should we wait for rate limits if necessary?
.json: Should output be in JSON mode (default: FALSE).
.min_tokens_reset: How many tokens should be remaining to wait until we wait for token reset?
.stream: Stream back the response piece by piece (default: FALSE).
.dry_run: If TRUE, perform a dry run and return the request object.

Value

Returns an updated LLMMessage object.