fair point, I already had the gpt-oss:20b model loaded so that was the first one I tested with to make sure ollama was successfully talking to Claude.

But with my hardware, I can't really bigger models than the 20b-30b range.

Thoughts on gwen3-coder:30b?