Can I run Llama 3 8b instruct Q4_K_M at an adequate speed?

Thinking of getting one of these to keep a small AI constantly running ,anyone’s got any experience trying to do that with this hardware and software?