Gctf is a GPU-accelerated computer program designed for real-time, accurate, and robust determination of the contrast transfer function (CTF) in cryo electron microscopy (cryoEM). The program aims to maximize the cross-correlation between a simulated CTF and the power spectra of observed micrographs after background reduction. Key features include GPU acceleration, a fast '1-dimensional search plus 2-dimensional refinement' procedure, and novel diagnostic methods such as Equiphase Averaging (EFA) and self-consistency verification. Gctf can process thousands of micrographs within minutes using a single GPU card, and it provides accurate and stable CTF parameters for subsequent image processing. The program also includes local defocus refinement for each particle and movie frames, which significantly improves 3D reconstructions. Extensive testing on various datasets demonstrates the program's robustness, speed, and accuracy, making it a valuable tool for near-atomic resolution cryoEM reconstruction.Gctf is a GPU-accelerated computer program designed for real-time, accurate, and robust determination of the contrast transfer function (CTF) in cryo electron microscopy (cryoEM). The program aims to maximize the cross-correlation between a simulated CTF and the power spectra of observed micrographs after background reduction. Key features include GPU acceleration, a fast '1-dimensional search plus 2-dimensional refinement' procedure, and novel diagnostic methods such as Equiphase Averaging (EFA) and self-consistency verification. Gctf can process thousands of micrographs within minutes using a single GPU card, and it provides accurate and stable CTF parameters for subsequent image processing. The program also includes local defocus refinement for each particle and movie frames, which significantly improves 3D reconstructions. Extensive testing on various datasets demonstrates the program's robustness, speed, and accuracy, making it a valuable tool for near-atomic resolution cryoEM reconstruction.