The release of large-scale datasets for natural language inference has enabled several advancements in modeling techniques. However, much of this progress is restricted to English language models. To address this problem, researchers from Indiana University and the AI2 have released the Original chinese Natural Language Inference dataset, which includes about 56,000 annotated sentence pairs with annotations from native speakers specializing in linguistics. The corpus, which is part of the CLUE benchmark, follows the annotation protocol used for MNLI.