Sustainable use of insect resistance in crops require insect resistance management plans that may include a refuge to limit the spread of virulence to this resistance. However, without a loss of fitness associated with virulence, a refuge may not prevent virulence from becoming fixed within a population of parthenogenetically reproducing insects like aphids. Aphid-resistance in soybeans (i.e., Rag genes) prevent outbreaks of soybean aphid (Aphis glycines), yet four biotypes defined by their capacity to survive on aphid-resistant soybeans (e.g., biotype-2 survives on Rag1 soybean) are found in North America. Although fitness costs are reported for biotype-3 on aphid susceptible and Rag1 soybean, it is not clear if virulence to aphid resistance in general is associated with a decrease in fitness on aphid susceptible soybeans. In laboratory assays, we measured fitness costs for biotype 2, 3 and 4 on an aphid-susceptible soybean cultivar. In addition, we also observed negative cross-resistance for biotype-2 onRag3, and biotype-3 on Rag1 soybean. We utilized a simple deterministic, single-locus, four compartment genetic model to account for the impact of these findings on the frequency of virulence alleles. When a refuge of aphid susceptible was included within this model, fitness costs and negative cross-resistance delayed the increase of virulence alleles when virulence was inherited recessively or additively. If virulence were inherited additively, fitness costs decreased the frequency of virulence. Combined, these results suggest that a refuge may prevent virulent A. glycines biotypes from overcoming Rag genes if this aphid-resistance were used commercially in North America.