[ad_1]
\n<\/p>\n
Developers can now describe functions to These models have been fine-tuned to both detect when a function needs to be called (depending on the user\u2019s input) and to respond with JSON that adheres to the function signature. Function calling allows developers to more reliably get structured data back from the model. For example, developers can:<\/p>\n Convert queries such as \u201cEmail Anya to see if she wants to get coffee next Friday\u201d to a function call like Convert \u201cWho are my top ten customers this month?\u201d to an internal API call such as Define a function called These use cases are enabled by new API parameters in our gpt-4-0613<\/code> and gpt-3.5-turbo-0613<\/code>, and have the model intelligently choose to output a JSON object containing arguments to call those functions. This is a new way to more reliably connect GPT’s capabilities with external tools and APIs.<\/p>\n\n
send_email(to: string, body: string)<\/code>, or \u201cWhat\u2019s the weather like in Boston?\u201d to get_current_weather(location: string, unit: 'celsius' | 'fahrenheit')<\/code>.<\/p>\n\n
get_customers_by_revenue(start_date: string, end_date: string, limit: int)<\/code>, or \u201cHow many orders did Acme, Inc. place last month?\u201d to a SQL query using sql_query(query: string)<\/code>.<\/p>\n\n
extract_people_data(people: [{name: string, birthday: string, location: string}])<\/code>, to extract all people mentioned in a Wikipedia article.<\/p>\n\/v1\/chat\/completions<\/code> endpoint, functions<\/code> and function_call<\/code>, that allow developers to describe functions to the model via JSON Schema, and optionally ask it to call a specific function. Get started with our developer documentation<\/a> and add evals<\/a> if you find cases where function calling could be improved
<\/p>\n<\/div>\n<\/div>\n