An important assumption underlying psychobiological studies relating stress reactivity with disease risk is that individuals are characterized by stable response profiles that can be reliably assessed using acute psychophysiological stress testing. Previous research has mainly focused on the stability of cardiovascular, neuroendocrine, and cellular immune responses to repeated stressors, and less attention has been given to inflammatory and platelet responses. We therefore examined both average stability and individual test-retest stability of cardiovascular, neuroendocrine, hemostatic, inflammatory, and subjective responses to mental stress over two repeated stress sessions, four weeks apart. Ninety-one healthy, non-smoking men (mean age 33.2 years) completed a 3-min speech task followed by a 5-min mirror tracing task on two separate occasions. Blood samples were taken at baseline and 10 min after the stress tasks while cardiovascular activity, saliva samples, and subjective ratings were measured repeatedly. There was significant cardiovascular and cortisol activation to the stressors and stress-induced increases in plasma C-reactive protein, von Willebrand factor antigen, and platelet activation indexed by leukocyte-platelet aggregates. The magnitude of stress responses did not differ between sessions in any variable. Significant test-retest correlations between sessions were observed for baseline and stress values of all variables (r=0.47-0.74, p<.001), but reactivity (change scores) for C-reactive protein, von Willebrand factor, cortisol, and platelet activation were not significantly correlated. Our results demonstrate that the stress-induced responses did not habituate between sessions, though the small magnitude of acute inflammatory, cortisol, and platelet responses limits the test-retest reliability of stress reactivity assessments.