Comment by bdavbdav

I’m not sure what model I’d trust locally with anything meaningful in Openclaw. The smaller/simpler the model is, the greater the chance of fluff answers is.