ǝʌɐp on Nostr: I was just thinking very roughly that R1 is probably broadly comparable to whatever ...

I was just thinking very roughly that R1 is probably broadly comparable to whatever claude-code's model is (likely 3.7-sonnet?). But yeah a lot of people have had success with the qwen-coder models. I think 14b is about the limit of what I can run efficiently atm too.