[Terretta]

LM Studio on MacOS provides an estimate of whether a model will run on the GPU, also lets you partially offload.

The underlying CLI tools do this, the app makes it easier to see and manage.