What is your point about Qwen? Or is it just a general statement regarding LLM?

I did it the old school way with openai as the fallback.

All I'm saying is I'm excited to try Qwen to see if it out performs my gnarly algorithm.