First off, you would have to check the F-Stop of the lens system being used to begin with.
Without knowing that you'd know nothing as it is perhaps the most important parameter to start with, but as OdDuMet previously mentioned, not many manufactures specify this data for comparison.
In regards to microbolometer performance, there are trade-offs between spatial, spectral, and radiometric resolutions when designing one.
For high spatial resolution, the sensor has to have a small IFOV, however as the IFOV becomes smaller its ability to detect photon energy becomes smaller too. This leads to reduced radiometric resolution or the ability to detect fine energy differences.
If a 12 micron thermoresistor takes one 12 micron photon of thermal emissivity to activate it and a 38 micron thermoresistor takes four 12 micon photons to activate it, which one do you think will provide a more stable output without further software normalization?