I am trying to do some debugging around costs and Helicone- when searching for a specific request string:
https://www.helicone.ai/requests?filters=%22%25220%253A2%253Alabel%2522%22&page=5&page_size=100
I see that these requests began yesterday at 3:10 PM.

However, these requests have been going on for longer than that, we shipped this feature before November 2nd and have not made any changes to the LLM call.

We are seeing a large cost discrepancy between OpenAI and Helicone (~$300 vs ~$90) / day and I am worried that the requests are not being properly handled

I am no longer seeing cost estimates for my project. This was working until maybe a week ago?

It started around the same time I enabled cacheing on the project, but I'm not sure that it was the exact same time or they are related