Since the redefinition of the kelvin in 2019, new methods of primary thermometry have been considered to replace the currently agreed temperature scale in the very high and low temperature limits. These new methods should provide improved uncertainties and, most importantly, a more direct link to the definition of the kelvin. We present an experimental comparison of two such primary thermometers working in the mK region: the Coulomb Blockade Thermometer (CBT) and the Shot Noise Thermometer (SNT). Both thermometers measure temperature hinging only on the natural constants $k_B$ and $e$. Furthermore both of them are based on electron tunneling current and, thereby, need only electrical measurements, enhancing the practicality. CBT and the SNT are inter-compared in a range of 20 mK to 235 mK. The results show that the agreement of SNT and CBT is approximately within 2.5~% in this range. Basic measurement uncertainty is analyzed and, additionally we show that uncertainty in the measurement frequency can cause significant error to temperature measurement of the SNT at low temperatures where finite frequency plays a role.