at the moment, I think the best you can do is qwen3-coder:30b -- it works, and it's nice to get some fully-local llm coding up and running, but you'll quickly realize that you've long tasted the sweet forbidden nectar that is hosted llms. unfortunately.
No comments yet.