It's more likely to work. Most LLM runners are meant to work with any model, which means there are all kinds of ways you might misconfigure them in a way that causes function tooling not to work, or performance to be less than you would like.

DwarfStar's selling point is that it only supports a small set of carefully chosen models, but it supports them really well.