Purpose Anti-Müllerian hormone (AMH) is used as a marker for ovarian reserve. Since 2011, the standard test for AMH has been the Beckman Coulter Generation (Gen) II assay. However, in July 2013, the protocol was revised due to falsely low readings. The aim of this study was to compare AMH levels measured with the original and revised Gen II assay and to establish a fertile female reference range for the revised protocol. Methods Serum AMH levels were measured for 492 natural conception first trimester pregnant women using the original and revised Gen II assay. Results The original protocol significantly underestimated AMH levels compared with the revised protocol (p<0.001), the median being 8.4 and 14.2 pmol/L, respectively. In all samples with detectable AMH levels, the revised protocol yielded a higher concentration compared with the original protocol, the magnitude shift ranging from 3.4 to 283.3 % (median 68.0 %). AMH levels measured with the revised protocol were collated to generate an age-specific reference range, with median levels peaking at 27 years then declining with advancing age. The median AMH concentration for ages 20-24 was 17.3 pmol/L, ages 25-29 was 20.5 pmol/L, ages 30-34 was 17.8 pmol/L, ages 35-39 was 10.8 pmol/L, and ages 40-44 was 6.1 pmol/L. Conclusions Our study demonstrated that the original Gen II assay significantly underestimated AMH levels, suggesting caution is required when interpreting literature and testing results achieved with this assay. We also established the revised Gen II assay reference range for AMH in women with unassisted proven fertility.