It uses an ADF4360 synth, driven by what looks like a standard osc module. The frequency was surprisingly accurate - ISTR less than 0.5ppm off my rubidium standard
Mine was also quite good for accuracy and stability, compared against GPS standard. Plot below shows 4 days or so worth of stability as measured against an HP3816A GPS frequency standard. You can see the day/night temperature change in my basement lab. Not sure what caused the abrupt shift on the right hand side of the plot - GPS may have had a lock interruption.