Apparently, it's really hard to measure biotin levels in food accurately.
From "Determination of the biotin content of select foods using accurate
and sensitive HPLC/avidin binding" (2004, emphasis added):
Total biotin content of 87 foods was determined using acid
hydrolysis and the HPLC/avidin-binding assay. These values are consistent with published values
in that meat, fish, poultry, egg, dairy, and some vegetables are relatively rich sources of biotin.
However, these biotin values disagreed substantially with published values for many foods. Assay
values varied between 247 times greater than published values for a given food to as much as 36%
less than the published biotin value. Among 51 foods assayed for which published values were
available, only seven agreed within analytical variability (720%). We conclude that published values
for biotin content of foods are likely to be inaccurate.