QIMMA قِمّة ⛰: A Quality-First Arabic LLM Leaderboard

The QIMMA leaderboard has been released to evaluate the performance of Arabic language models, emphasizing quality metrics over traditional benchmarks. It includes a range of models with varying sizes and architectures, providing detailed performance scores across tasks such as text generation and comprehension. This initiative is significant for practitioners as it offers a standardized framework for assessing Arabic LLMs, facilitating better model selection and development in Arabic NLP applications.