Background. Suicide capability (fearlessness about death [FAD], preparation, and perceived capability [PC]) is presumed to be static, or to increase with exposure to provocative events. However, tests of this assumption have primarily used non-clinical samples. We examined short-term changes in capability among patients in a partial hospital program. Methods. We enrolled 296 adult patients (186 [62.7%] female; Mage=36.13, SDage=14.75) who completed diagnostic interviews and questionnaires assessing demographics, capability, and suicide ideation (SI) at intake. Capability was re-assessed twice during treatment and again at discharge, and SI was re-assessed at discharge. We used latent growth modeling (LGM) to quantify changes in capability. Results. FAD, preparation, and PC decreased from intake to discharge (ps<.001, ds=0.24-0.63) and unconditional LGM models suggested that all facets changed significantly throughout treatment. FAD was positively associated with preparation (r=0.57) and PC (r=0.25) at intake (ps<.001), but changes in FAD did not predict changes in other facets. Greater SI at intake was concurrently associated with higher capability (βs=0.26-1.83, ps<.01) and predicted steeper declines in preparation (β=-0.23) and PC (β=-0.04), ps=.04. Finally, higher intake preparation predicted more severe SI at discharge (β=0.16, p<.001). Limitations. There was no control group and there was unequal spacing between in-treatment assessments among participants. Results for SI may have been impacted by floor effects. Conclusions. Results suggest that aspects of capability can change over a short time. Conceptually related facets of capability did not change together; thus, future investigations of the short-term dynamics of capability should not treat it as a unitary construct.