I use CoreML for models designed for other platforms. I port the models to it but it works for that without much trouble.

MLX is not for end user deployment.