Wrong model being used?

We are testing cursor with our OpenAI key and it’s set to be GTP-4, and in the billing we are seeing InstructGPT is this expected ?

Could you send a screenshot of how your chat is configured? Does it say “gpt-4” in the bottom left of the chat bubble?

Note that you may see some billed items for gpt-3.5 also — for example, we use it for summarization and building up the context for gpt-4 in some places.

When you say InstructGPT, which model do you mean? gpt-3.5-turbo-instruct? We are basically not using that model at all yet.

This is not happening this month, I’ll keep checking and report if it happens again.

1 Like