You can probably run it locally with llama.cpp using CPU only, but it will be slow. I have a couple year old laptop with a RTX 3060 and it runs pretty well spli

You can probably run it locally with llama.cpp using CPU only, but it will be slow. I have a couple year old laptop with a RTX 3060 and it runs pretty well split across the CPU and GPU.