The tuning of the ACME V1 atmosphere model at high-resolution (0.25-degree) evolves with each major stage of the tuning at 1-degree resolution. Physical parameterizations in the atmospheric model are sensitive to resolution. High-resolution baseline simulations based on well-tuned model at low resolution tend to always exhibit significant biases that call for further tuning specifically for high-resolution configuration. Tuning at high-resolution is computationally expensive. To reduce computational cost and meanwhile improve turnaround time, the high-resolution tuning strategy emphasizes the use of short-term hindcasts based on the Cloud Associated Parameterization Testbed (CAPT) framework to poll the tuning response. The parameter choice for the CAPT-based tuning takes guidance from the Perturbed Parameter Ensemble (PPE) simulations at low resolution. Only the best performing CAPT tuning is then used for conventional tuning with AMIP-like long simulations. The iterative combined approach is found to be effective, leading to improved overall simulations with only a small number of expensive long simulations. Preliminary results in comparisons with the low resolution simulations and observations will be shown. Emphasis will be placed on biases exist at low resolution but amplified or reduced at high-resolution. Challenges for further tuning will be discussed, including computational burden and stubborn biases unique in high-resolution simulations.