ArabicCulturalQA: leaderboard
QCRI benchmark of Arabic cultural knowledge: multiple-choice questions across Modern Standard Arabic, Egyptian, Levantine, Maghrebi and Gulf dialects plus English; zero-shot MCQ accuracy averaged over the six splits.
Metric: MCQ Accuracy Avg (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 | 79.1 | #258 |
| 2 | GPT-5 | 78.4 | #105 |
| 3 | ALLaM-7B-Instruct-preview | 64.7 | #724 |
| 4 | Fanar-1-9B-Instruct | 62.6 | #635 |
| 5 | Falcon3-10B-Instruct | 48.5 | #1147 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=arabicculturalqa · How It Works · Data refreshed daily, snapshot 2026-11-05.