suggestions on GPT-3 api Token usage.

Hello everyone.
When creating an app that summarizes bulky content (e.g an article) using GPT-3 model, Is it possible to exclude the content to be summarized from the prompt?

The challenge currently is, because i need to send the whole content along with the prompt, it significantly increases the amount of token used for each api call from less than 200 tokens to over 1000 tokens.

has anyone experienced such before or have knowledge of how i can resolve this?