QIMMA قِمّة ⛰: A Quality-First Arabic LLM Leaderboard

QIMMA is a quality-validated Arabic LLM leaderboard that validates benchmark quality before evaluating models, addressing fragmentation in Arabic NLP. The authors discovered systematic quality issues in widely-used Arabic benchmarks including translation artifacts and annotation inconsistencies. The leaderboard consolidates 52,000+ samples from 14 benchmarks across 7 domains with 99% native Arabic content.