What is the actual depth of the 4th soil layer in ERA5?

Datasets and Usage C3S - Datasets and Usage
1

I’m seeing conflicting information regarding the depth of the 4th soil layer in ERA5.
Using cdo zaxisdes on grib files downloaded either from Mars or from the ’ ERA5 hourly data on single levels from 1940 to present’ interface results in:

#
# zaxisID 5
#
zaxistype = depth_below_land
size      = 1
name      = depth
longname  = "depth_below_land"
units     = "cm"
levels    = 100 
lbounds   = 100 
ubounds   = 255

or when I use cdo sinfo:

     2 : depth_below_land         : levels=1
                            depth : 0 cm
                           bounds : 0-7 cm
     3 : depth_below_land         : levels=1
                            depth : 7 cm
                           bounds : 7-28 cm
     4 : depth_below_land         : levels=1
                            depth : 28 cm
                           bounds : 28-100 cm
     5 : depth_below_land         : levels=1
                            depth : 100 cm
                           bounds : 100-255 cm

This indicates the 4th layer extends until 255 cm below the ground, while the documentation states it is 289 cm. Can anyone clarify which value is correct and why there might be this discrepancy?