Fact · Languages and accents
The Open Universal Arabic ASR Leaderboard is a continuous benchmark of open-source Arabic speech models across several multi-dialect datasets.
Checked against its source on October 17, 2026
In the source's own words
“a continuous benchmark project for open-source general Arabic ASR models across various multi-dialect datasets”
Open Universal Arabic ASR Leaderboard, Yingzhi Wang et al., arXiv, 2024, Abstract.
Open the original- What it measures
- The leaderboard is one of the few public places where Arabic models are compared across dialects, and it reports robustness, speaker adaptation and memory use alongside accuracy.
- What it does not show
- Open-source models only. Commercial services are not in the leaderboard, and the rankings change as new models are added.