Accuracy is the most fundamental and commonly used evaluation metric for classification and judgment tasks. It calculates the proportion of samples where the predicted results exactly match the true labels, intuitively reflecting the model's overall correctness. It is typically reported as a percentage or a score between 0 and 1, with higher values indicating better performance. This metric is reliable when classes are balanced; for severely imbalanced classes, it is recommended to use it in conjunction with Macro-F1, MCC, or AUROC. This metric is widely used in related benchmarks and papers, facilitating horizontal comparisons across different methods or model versions.