We present results of our investigation of the radio intrinsic brightness temperatures of compact radio jets. The intrinsic brightness temperatures of about 100 compact radio jets at 2, 5, 8, 15, and 86 GHz are estimated based on large VLBI surveys conducted in 2001-2003 (or in 1996 for the 5 GHz sample). The multi-frequency intrinsic brightness temperatures of the sample of jets are determined by a statistical method relating the observed brightness temperatures with the maximal apparent jet speeds, assuming one representative intrinsic brightness temperature for a sample of jets at each observing frequency. By investigating the observed brightness temperatures at 15 GHz in multiple epochs, we find that the determination of the intrinsic brightness temperature for our sample is affected by the flux density variability of individual jets at time scales of a few years. This implies that it is important to use contemporaneous VLBI observations for the multi-frequency analysis of intrinsic brightness temperatures. Since our analysis is based on the VLBI observations conducted in 2001-2003, the results are not strongly affected by the flux density variability. We find that the intrinsic brightness temperature $T_0$ increases as $T_0{propto}{ u}^{xi}_{obs}$ with ${xi}=0.7$ below a critical frequency ${ u}_c{approx}9GHz$ where the energy loss begins to dominate the emission. Above ${ u}_c$, $T_0$ decreases with ${xi}=-1.2$, supporting the decelerating jet model or particle cascade model. We also find that the peak value of $T_0{approx}3.4{ imes}10^{10}$ K is close to the equipartition temperature, implying that the VLBI cores observable at 2-86 GHz may be representing jet regions where the magnetic field energy dominates the total energy in jets.