I have 8GB RAM. So I am able to run 7B 4bit at maximum.

I used GGUF + llama.cpp.

Will try out Ollama.

Thanks for the info!