Fast Requests Calculation Update

I wonder if there are packages for unlimited fast requests as such would come in handy when intensive work periods of my project. I would not mind paying for additional 60-80 USD on top of the pro pack.

Anyway, I don’t know how this calculation works though only 5 days in my pro-pack membership cycle has already occurred additional 20$ charge.

Sonet 3.7 is only deducting query counts but stopping without returning premium query results. So far, I’ve lost about 20 queries. Also, for the diff issue, quick queries are being deducted, but there are no actual modifications.


Just now…

yeah man! I’m encountering the same issue for like 4h now…

Every time I retry or resume, I get taxed…

I estimate that I lost 250-300 credits during this period, but I have not been refunded anything

I see its only if it fails but if I get this message

“We’re experiencing high demand for Claude 3.7 Sonnet right now. Please switch to the ‘default’ model, Claude 3.5 sonnet, another model, or try again in a few moments”

It still charges on fail and I have tested this over weekend and at least burnt 100 tokens that hit high token demand with few fails and it still charged.

It generally fails once and that gets refunded but any request after that is charged especially if it fails during High token demand