The aim of this study was to evaluate large-bean coffee genotypes from the Big Coffee VL group in terms of bean size and the percentage of post-harvest defects, according to the official Brazilian classification. The experiment was carried out in the Coffee Crop Sector of the Department of Agriculture at the Federal University of Lavras, where the progenies with the greatest production potential were selected, known as "elite" progenies. The experimental design was the Completely Randomized Design. Twelve progenies were selected, arranged in 6 replicates. The samples consisted of 300 grams of fruit from each progeny, which were classified according to grain size: large grains were those retained on the 17 (round sieve) and 13 (oblong sieve) sieves; medium grains were those retained on the 15 (round sieve) and 10 (oblong sieve) sieves and small grains were those retained at the bottom of the sieves. The percentage of mocha grains was obtained from the sum of all the sieves for mocha grains 10 and above. The data was submitted to analysis of variance and grouped using the Scott-Knott test at 5% significance. It was observed that there is genetic variability between the progenies that are segregating for the grain size characteristic, with some progenies standing out for their superior performance: M20, M4, G16 and G17. In terms of defect analysis, P14, P23, P32, M4, M5, M11, M20, G9 and G16 showed the lowest percentage of bored fruit. As the selected progenies are still segregating, further studies could be carried out.