rasterized, dpi, bbox_inches='tight' and axins mess up the plot in a pdf

When I run the code below, the plot in the PDF is messed up. The content of the plot and the colorbar are shifted down to the left, as shown in the attached screenshot of the PDF.

import matplotlib.pyplot as plt
import numpy as np
import os
from mpl_toolkits.axes_grid1.inset_locator import inset_axes

fig, ax = plt.subplots()

# create random image
image = np.arange(100).reshape((10,10))
xaxis = np.arange(10)
yaxis = np.arange(100)[::10]

im = ax.pcolormesh(xaxis, yaxis, image, shading='auto', rasterized=True)

axins = inset_axes(ax,
                    width=0.06,
                    height='100%',
                    loc='center right',
                    borderpad=-0.75
                    )
cbar = fig.colorbar(im, cax=axins, orientation='vertical')

filename = 'test.pdf'
fig.savefig(filename,
            dpi=300,
            bbox_inches='tight')

This problem does not occur if I do any of the following:

• Remove the dpi=300 keyword in fig.savefig() (although there is a minimal misplacing when zooming far in)
• Remove the keyword bbox_inches='tight' in fig.savefig()
• Remove the rasterized=True in ax.pcolormesh()
• Don't add the colorbar with inset_axes().

My matplotlib version is 3.9.2.

Any help is greatly appreciated!

This is a known bug that will be fixed in the upcoming Matplotlib v3.11 release. However, there are two different inset_axes functions within Matplotlib. The more modern axes method one does not have this problem.

axins = ax.inset_axes([1.01, 0, 0.02, 1])

With Matplotlib v3.10.5, this gives me: