BLEU (Bilingual Evaluation Understudy), proposed by Kishore Papineni et al. from IBM Research in 2002, is one of the most classic automatic evaluation metrics in the fields of machine translation and text generation. It calculates the n-gram precision (typically 1 to 4 grams) between the candidate and reference texts, multiplied by a brevity penalty factor to derive a comprehensive score. BLEU-4 considers up to 4-grams. The score ranges from 0 to 100, typically reported as a percentage or a score between 0 and 1, where higher values indicate better performance and greater overlap with the reference at the phrase level. It is often used in conjunction with semantic metrics to compensate for its insensitivity to synonymous expressions.