We estimate productivity growth due to genetic improvement in soybeans using a unique yield measure derived from Illinois State Agricultural Experiment Station (SAES) data. Seed companies can pay the SAES to grow their soybean varieties under experimental conditions and publish the results for the public. In the Illinois soybean trials, three public sector varieties have been planted in each year: Williams, Jack, and Dwight. By comparing each year’s varieties to these public varieties, we extract a measure of yield improvement due to genetic improvement that is not confounded by weather and management. We find results in line with other heritage studies, about a 30 kg/ha improvement each year in maturity groups two and three. Unlike the literature, we do not find uniform growth due to GE varieties across maturity groups: maturity group three soybeans grew at the same rate, whether conventional or GE. We also find evidence of a slow down in genetic improvement in the period 2009 to 2015. Our study is the first to estimate genetic improvement in soybeans after 2010 and the first study that is able to analyze the variance of genetic improvement within a given year.