Character limit (response) for the GPT-3.5 API

What is the character limit (response) for the GPT-3.5 API, and what are the pricing plans to increase that limit

1 Like

AI models have what is called a context length.

The context length is used for receiving the input that you provide, and then the remaining space is where tokens continue to be generated to create an output, which can be up to the limit of this memory.

The input includes everything you send the model: system message, function definitions and function returns, chat history, and and the latest input.

Context lengths:
gpt-3.5-turbo: 4097 tokens
gpt-3.5-turbo-16k: 16385 tokens ( 2x cost for both input and output)

In terms of characters, 3-4 characters of English are about one token. This can drop all the way to 0.4 characters per token in Chinese.