Unpredictable environmental changes displace individuals from homeostasis and elicit a stress response. In vertebrates, the stress response is mediated mainly by glucocorticoids (GCs) which initiate physiological changes while minimizing allostatic load. Individuals and species vary consistently in baseline and stress-induced GC levels and the speed with which GC levels can be upregulated or downregulated, but the extent to which variation in hormone regulation influences baseline and stress-induced GC levels is unclear. Using mathematical modeling, we tested how GC regulation rate, frequencies and durations of acute stressors, fitness functions, and allostatic load affect GC levels during control and acute stress periods. As GC regulation rate slows, baseline and acute stress-induced GC levels become more similar. When the speed of up- and down-regulation decreased, hormone levels became more linked to anticipated future conditions to avoid fitness costs of mismatching a new environmental state. When fitness was more tightly linked to hormone levels during acute stress periods than during control states, the speed of upregulation influenced optimal hormone levels more than downregulation rate. More frequent acute stressors caused baseline and acute stress-induced GC levels to converge. With allostatic overload costs included, predicted GC levels were lower and were more dependent on the frequency of past acute stressors. Our results show the value of optimality modeling to study the hormonal response to stressors and suggest GC levels depend on past and anticipated future environmental states, and individual differences in hormone regulation.