I have a Replit hosted AI agent with an OpenAI API assistant, for which I am consistently seeing unusual billing amounts that do not match the pricing listed on OpenAI website, see the sample conversation attached.
This seems really odd and I am not able to get to the bottom of why I can never forecast anything related to token consumption and API token usage even though its listed specifically on its website.
#Why does OpenAI API pricing vs actual token consumption not matching reality (EVER)?
tiny nebula
steady apex
I see your sample conversation but nothing in there indicates what charges you are actually seeing from this conversation