IndicParam is a graduate-level benchmark designed to evaluate Large Language Models (LLMs) on their understanding of low- and extremely low-resource Indic languages.The dataset contains 13,207 multiple-choice questions (MCQs) across 11 Indic languages, plus a separate Sanskrit–English code-mixed set, all sourced from official UGC-NET language question papers and answer keys.

	Supported Tasks… See the full description on the dataset page: https://huggingface.co/datasets/bharatgenai/IndicParam.