That card's been my mainstay since 2018. Solid device! Can run 7B LLMs under quantization. Some 13B also work but kiss the GUI rendering good bye during iterations - vram space will be tight. XD