I often fork chats to try the same prompt on different models. The problem is that the cost estimate from the initial chat carries over to the forked chat so that I don't really have a good idea of the cost of the forked chat.

For example, I may engage in a chat with Claude 3 Opus and incur an estimated cost of $2.50. I then fork the chat and use GPT-4o. The estimated cost of the forked chat is not reset so it begins at $2.50 even though this amount was spent on the earlier chat and not on the forked chat. I continue the original chat until the estimated cost reaches $4. I also continue the forked chat until the estimated cost reaches $3.

The next question is what is the estimated amount I spent on the two chats? The answer is $4.50 ($4 on the first chat and $0.50 on the second chat). However, this is extremely difficult to figure out based on the estimated costs. Instead of simply adding them, I must remember when I forked the chat and how much of the estimated cost from the original chat carried over to the forked chat. This quickly becomes an impossible task. I would rather have the estimated cost reset when a chat is forked so that I can simply add estimated costs of each chat to figure out the total estimated cost.