XIFBench: Evaluating Large Language Models on Multilingual Instruction Following

Min Zhang (Harbin Institute of Technology, Shenzhen) · Yunfei Long (Queen Mary, University of London) · Kehai Chen (Harbin Institute of Technology (Shenzhen)) · Xuefeng Bai (Harbin Institute of Technology) · Zhenyu Li (Harbin Institute of Technology (Shenzhen)) · Yaoyin Zhang (Harbin Institute of Technology, Shenzhen) · Xuchen Wei (Harbin Institute of Technology, Shenzhen) · Juntao Li (Soochow University, China)

constraint categoriesconstraint-based benchmarkconstraint-level translation validationcultural accessibility annotationdetailed insightsfine-grained analysisinstruction complexitylinguistic contextsmultilingual evaluationmultilingual instruction-followingperformance disparitiesrequirement-based evaluationresource levelssemantic anchors

Large Language Models (LLMs) have demonstrated remarkable instruction-following capabilities across various applications. However, their performance in multilingual settings lacks systematic investigation, with existing evaluations lacking fine-grained constraint analysis across diverse linguistic contexts. We introduce **XIFBench**, a comprehensive constraint-based benchmark for evaluating multilingual instruction-following abilities of LLMs, comprising 558 instructions with 0-5 additional constraints across five categories (*Content*, *Style*, *Situation*, *Format*, and *Numerical*) in six languages spanning different resource levels. To support reliable and consistent cross-lingual evaluation, we implement three methodological innovations: cultural accessibility annotation, constraint-level translation validation, and requirement-based evaluation using English requirements as semantic anchors across languages. Extensive experiments with various LLMs not only quantify performance disparities across resource levels but also provide detailed insights into how language resources, constraint categories, instruction complexity, and cultural specificity influence multilingual instruction-following. Our code and data are available at https://github.com/zhenyuli801/XIFBench.