Background/Objectives: Programs combining resistance training with a concurrent cognitive demand are often presented as superior to resistance training alone, although the contribution of the cognitive component has never been quantified. This review quantified the effect of dual-task strength training (DTST) on dual-task performance, lower-limb strength, functional mobility, cognition and balance in older adults, relative to the same resistance program performed without the cognitive task. Methods: Three databases and Google Scholar were searched to March 2026 following PRISMA 2020, with eligibility requiring the cognitive task to be executed simultaneously with resistance exercise. Risk of bias was assessed with RoB 2 and ROBINS-I and certainty with GRADE, and random-effects models were fitted by restricted maximum likelihood. Results: Fourteen studies were included, and seven randomized trials with an active comparator entered the meta-analyses. Dual-task performance favored DTST (g = 0.53, 95% CI 0.25 to 0.81; 203 participants; I² = 0%; very low certainty), as did cognition (g = 0.91, 95% CI 0.34 to 1.49; 143 participants; I² = 7%; very low certainty), whereas lower-limb strength did not differ (g = 0.20, 95% CI −0.14 to 0.54; 135 participants; low certainty). The single-task Timed Up and Go favored DTST by 0.48 s (95% CI 0.09 to 0.88; 179 participants; low certainty), with a point estimate below the change regarded as clinically meaningful. Balance findings diverged across non-interchangeable instruments (I² = 87%), and only two studies monitored falls prospectively. Conclusions: DTST may make little or no difference to lower-limb strength and may slightly improve single-task mobility, whereas its effects on performance under cognitive constraint and on cognition are very uncertain despite favorable point estimates. Part of the dual-task gain may reflect practice on the task used at testing, and the evidence, of low to very low certainty, does not support DTST as a replacement for conventional resistance training.