This paper investigates the iterates b̂^1, …, b̂^T obtained from iterative algorithms in high-dimensional linear regression problems, in the regime where the feature dimension p is comparable with the sample size n, i.e., p ≍ n.  The analysis and proposed estimators are applicable to Gradient Descent (GD), proximal GD and their accelerated variants such as Fast Iterative Soft-Thresholding (FISTA).  The paper proposes novel estimators for the generalization error of the iterate b̂^t for any fixed iteration t along the trajectory. These estimators are proved to be √n-consistent under Gaussian designs.  Applications to early-stopping are provided: when the generalization error of the iterates is a U-shape function of the iteration t, the estimates allow to select from the data an iteration t̂ that achieves the smallest generalization error along the trajectory.  Additionally, we provide a technique for developing debiasing corrections and valid confidence intervals for the components of the true coefficient vector from the iterate b̂^t at any finite iteration t.