How do I get this to work exactly

[RandomTurtle]

I tried multiple models and they all just talk back or hallucinate. How do I get this to actually code? I want to run it locally, so I try out Qwen2.5-coder and it hallucinates a bunch of json text into the screen. llama acts like a clueless silly chatbot, mistral keeps asking me to 'provide more details' on what I need help with when I just ask it 'list the files in the current directory' I just want to be able to vibe without being cut off but these models are all loopy

[acoliver]

I have a new series I started for this: https://m.youtube.com/playlist?list=PL2HVRBhH2-QPGR_zRsFI6wq5kNorM8IYY

Goal is to show exactly that!