A collaborative initiative launched by the academy in partnership with leading academic and government institutions across the Middle East. The initiative focuses on developing a specialized platform to evaluate Arabic computing technologies, including the creation and preparation of essential test datasets for assessing the performance of large language models (LLMs) in various natural language processing (NLP) tasks in Arabic. The BALSAM index offers a unified reference that enables developers, researchers, and institutions to better understand the strengths and weaknesses of language models by evaluating their performance in key tasks such as grammar checking, text generation, and content comprehension.

Its Objectives:
•    Measure the performance of large language models in Arabic language processing tasks.
•    Establish global standards for assessing the performance of Arabic language tasks.
•    Foster research collaboration among research groups interested in Arabic language processing.
•    Promote the development of standardized datasets that contribute to performance evaluation.

For more information: