HBSP simulation of Gilbert Cell Mixer

I design a Gilbert Cell mixer with the following port detail:

LO: 4.9 GHz (from VCO)

Port1 (RF) : 2.5 GHz

Port2 (IF) : 2.4 GHz

and I'm running hb/hbsp analys

here I want to find out how many dB that signal at 2.5GHz leaking from RF port is attenuated at IF port - RF-to-IF isolation.

So I plotted S21 (dB)

I know that S-parameter does not capture the frequency change but here the isolation is measured at a specific frequency (2.5GHz) 

To double check, I also plotted the hb spectral result to see the isolation and found that the 2 number were closed ( some dB different)

So I want to know if my S21 hbsp result is consider valid in this case

Thank you so much.