I don't use the assistant a lot, compared to some other users. Glancing at my billing settings, I was surprised to see how much usage I had this month:
If I compare August to the past two months - before that, there is just 0.00 for AI cost in my billing - it comes out to about 20x the previous cost per token. I understand that different models have different costs. I mostly use Ki, and I understand that it decides which model to use when I submit a query.
I don't keep all my queries, but this is the one that used the most tokens, reading a long PDF file. The cost for this is minimal.
Since I don't use the assistant a lot, I think there must be a problem with costing. I'm surprised to reach half my quota this month, compared to other months.
.
