In neuroendocrine tumors (NETs), proliferation markers, especially Ki-67, have become increasingly important. This study was designed to examine the reproducibility of Ki-67 for use in the current classification of NETs. A retrospectively assembled integrated database with prospectively collected data of patients undergoing multidisciplinary management for NETs from 2000 to 2009 was analyzed. Original pathology was reviewed to reassess Ki-67 values. Ki-67 was then categorized to grades G1 (≤2 %), G2 (3-20 %), or G3 (>20 %) according to the European Neuroendocrine Tumor Society (ENETS) guidelines and the 2010 World Health Organization (WHO) classification. Original Ki-67 values were compared to reviewed values. All statistical analyses were carried out using SAS 9.1.3. A total of 184 patients were included of which 48 % were male. The most common primary NET site was the small bowel, in 27 %. On pathology review, there was 94 % agreement for G1, with 4 % of cases upgraded at review to G2 and 2 % of cases upgraded to G3. For G2, there was 94 % agreement, with 6 % of cases downgraded to G1 and 0 % upgraded. For G3, there was 90 % agreement, with 10 % of cases downgraded to G2 and none to G1 (kappa = 0.89). Ki-67 is a proliferative marker for NETs that is highly reproducible when used to grade tumors according to ENETS and WHO categories. The high inter-institutional reliability in the determination of tumor grade as assessed by Ki-67 makes it a reliable tool in the assessment of patients with NETs.