Comment by wiremine

I ran a fairly large experiment last week, and the token usage wasn't bad at all. What softs of use cases are you seeing large token usage by GLM 5.2?