Why velocity azimuths in the reconstruction snapshots do not match those in the Kinematics Tool?

Dear Folks
Trying to use Gplates, I often see that the velocity azimuths in the reconstruction images are different from the azimuths in the Kinematics Tool. I use GPlates 2.5.0_win64 and dataset for the Plate Motion Model of Zahirovic et al, 2022.
The example of such a discrepancy is demonstrated in the attached images for the reconstruction of the Pacific plate (ID 901) at 48 and 47 Ma.
Anchored Plate ID = 0.
Time step is set to 1 Ma
In the reconstruction images we see azimuths ~330 degrees at 48 Ma and ~270 at 47 Ma.
The Kinematics Tool gives azimuths ~5 degrees at 48 Ma and ~291 at 47 Ma (at Lat 60 Lon 150).
Which results are correct and why does this discrepancy occur?
Is this problem related to the dataset or to the functioning of the software?
I will greatly appreciate any advice/suggestion…