Biopython SASA looks wrong and disagrees with and Pymol SASA

Question

I'm trying to calculate solvent accessible surface area using Biopython.

However, the values from biopython do not make sense when looking at the crystal model visually in Pymol. Furthermore, when I calculate the SASA using Pymol, those values match well visually, but do not agree with the biopython values.

Why is biopython producing supposedly wrong SASA values?

Here are the values given by Biopython (~180 is max):

[('GLY', 3.4), ('ILE', 0.0), ('VAL', 0.0), ('GLU', 4.25), ('GLN', 0.85), ('CYS', 0.0), ('CYS', 0.0), ('THR', 45.57), ('SER', 55.2), ('ILE', 42.79)]

You can see that Gly1 has a tiny value. However, looking at the model visually (see picture, surface in blue), Gly 1 has a lot of exposed surface.

Also, the calculated value using pymol is 67.44. This shows some exposure. Why is the biopython code so different and seemingly wrong?

Biopython code:

from Bio.PDB.SASA import ShrakeRupley

from Bio.PDB.MMCIFParser import MMCIFParser

parser = MMCIFParser(QUIET=True)
structure = parser.get_structure("HELLO", "insulin.cif")

sr = ShrakeRupley()

sr.compute(structure[0], level="R")

my_list = []
for chain in structure[0]:
    for res in chain:
        my_list.append((res.get_resname(),round(res.sasa,2)))

print(my_list[0:10])