numbers = {n: eisenstein_series_qexp(4,101,
normalization='constant')[n]
for n in [0..100]}An exact value -- an integer, a rational or a polynomial -- so there is no precision to choose and nothing to be wrong about.