This is the output from my NanoVNA as measured on an HP 8568B spectrum analyzer in Max Hold mode (50 Ohm input impedance). See attached photos.
~10 MHz -13 dBm
~100 MHz -9 dBm
~300 MHz -9.60 dBm
~600 MHz -20.6 dBm
~900 MHz -23.2 dBm
Therefore, there is an approximately 10 to 13 dBm drop in output level going above 300 MHz which is understandable due to the higher frequencies being derived from harmonics. What I am puzzled by is why the output is low (in the -13 dBm range) for frequencies below 100 MHz?