The $500/hr figure is for a 1965 vintage electrostatic "tandem" accelerator with a 10MV max terminal voltage. We mostly make beams of protons, deuterons, and helium. A common use case is making tightly pulsed deuteron beam, and then hitting a gas target filled with deuterium to make neutrons. The outgoing neutrons will also be pulsed and have a well defined energy spectrum, which makes them great for doing experiments with neutrons. At these energies they are non-relativistic, so you can measure their energy through time-of-flight.

The more expensive facility is for making beams of gamma rays. It uses a free electron laser built into a ~1GeV electron storage ring to make ultraviolet laser light in a building-length laser cavity. The laser light then collides directly with an electron bunch and is scattered up to gamma ray energy. It then passes through the laser cavity mirror and on to the target rooms. This is a unique kind of facility because it produces gamma beams with high resolution. It's complimentary to bremstralung gamma beams (often made by shooting electrons directly into a piece of diamond to make them wiggle).

We typically use the gamma rays to study photonuclear reactions. Having an electromagnetic probe instead of another hadron hitting your target allows for alot of simplifications on the theory end and helps theorists to generate testable predictions with less uncertainties attached. It also probes slightly different physics.