This depends on the structure of the device especially the location of contacts. In most cases, the charge transport is mainly 2D, either in lateral or vertical direction (when looking at the cross-section of the device) and no major charge transport is expected in light propagation direction so a 2D simulation can fairly represent the charge transport. If that is not the case, then a 3D simulation might be necessary but it might be very time consuming.
I’m not sure how this calculation can be done but one thing that is for sure is that a transient simulation for bandwidth calculation (as demonstrated in VPD example) can capture the effect of carrier transit time limit on bandwidth since it is a time dependent simulation.
This is because this particular device is most likely not limited by its RC bandwidth due to its size but even if that is not the case, the transient or SSAC simulation will also be able to capture the RC limited bandwidth (but not separately) meaning that it will be included in the bandwidth reported by the methods if it is the dominant factor. But usually RC becomes dominant when the device is big (large C) or there is an external load resistor (large R). This external R can only be considered in a transient simulation currently. If you want the RC bandwidth separately, then the approach used in UTC example should work.