METEOR is an automatic evaluation metric for machine translation, designed to mitigate perceived weaknesses in BLEU. METEOR scores machine translation hypotheses by aligning them to reference translations, much like BLEU does.