Abstract INTRODUCTION With anti‐amyloid beta (Aβ) therapies approved for Alzheimer's disease (AD), surrogate biomarkers are needed to monitor clinical treatment efficacy. Therefore, we systematically compared longitudinal changes in A/T/N biomarkers (amyloid‐positron emission tomography [PET], tau‐PET, plasma phosphorylated tau at threonine 217 [p‐tau 217 ], and magnetic resonance imaging) for tracking cognitive changes. METHODS We analyzed longitudinal biomarker and cognitive change rates from the Alzheimer's Disease Neuroimaging Initiative ( N = 141) and Anti‐Amyloid Treatment in Asymptomatic Alzheimer's (A4) and Longitudinal Evaluation of Amyloid Risk and Neurodegeneration (LEARN) ( N = 151), estimated using linear mixed models. Using linear models, we tested biomarker changes as predictors of cognitive changes, comparing predictive strengths across biomarkers using bootstrapping. RESULTS Tau‐PET, plasma p‐tau217, and cortical thickness changes accurately tracked change rates in Mini‐Mental State Examination, Alzheimer's Disease Assessment Scale‐Cognitive Subscale 13‐item version, Clinical Dementia Rating‐Sum of Boxes, and Preclinial Alzheimer Cognitive Composite scores. In contrast, amyloid‐PET change rates were not linked to cognitive changes. DISCUSSION Plasma p‐tau 217 offers a cost‐effective AD‐specific alternative to tau‐PET and could potentially be implemented for monitoring cognitive changes in AD trials, while amyloid‐PET lacks utility. Cortical thickness changes accurately track cognitive changes but may be confounded by pseudo‐atrophy in anti‐Aβ treatments. Highlights Longitudinal changes in tau‐PET, plasma p‐tau 217 , cortical thickness – but not amyloid‐PET – effectively track cognitive decline. Cortical thickness may be confounded by pseudo‐atrophy in anti‐Aβ trials. Plasma p‐tau 217 is a robust and cost‐effective alternative to tau‐PET as an AD‐specific surrogate biomarker for monitoring cognitive changes.