LongText-Bench

vision

LongText-Bench evaluates text-to-image models on their ability to accurately render long text passages within generated images. It includes English (EN) and Chinese (ZH) subsets to assess multilingual text rendering capabilities.