Are people doing image generation really using these models much? I've generated a lot of images, but I always use ComfyUI with local models and custom workflows. I only have 8GB of VRAM and I can easily do 1000s of images per day if I want to.