I have some interns starting next week, and I am thinking about just giving them credits on Openrouter. Pareto seems like a good way to keep them from using max model all the time. Is that a good use case? What's your experience with it?
I purchased a year for approximately $13/month. I get 4500 text requests per session. I've never been able to get near 25% of that number. Until today.
Data processing hundreds of math curriculum files with output:
* 346 files
* 45K lines of .md and json files
5-10 subagents running continuously.
I finally broke 25%.
I'm done with Z.ai. The abuse just never ends. They can have the last two weeks of my three-month sub. I don't care anymore.
It's like giving away your belongings to an abusive partner just so they'll leave.