I think this is less of an issue since these GPUs are not meant for the everyman, so basically the handful of server integrators can figure this out by themselv

I think this is less of an issue since these GPUs are not meant for the everyman, so basically the handful of server integrators can figure this out by themselves. And for your typical dev - they'll interact with the GPU through a cloud provider, where they can easily know that a G5 instance is newer than a G4 one.