Open Arabic LLM - Arabic MMLU HT Jurisprudence — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.3 70B Instruct | 78.7 |
| 2 | Qwen 2.5 72B Instruct | 78.7 |
| 3 | gemma-3-27B-pt | 77.78 |
| 4 | calme-2.2-qwen2.5-72B | 77.78 |
| 5 | Rombos-LLM-V2.5-Qwen-72B | 77.78 |
| 6 | Ultiima-72B | 77.78 |
| 7 | calme-2.1-qwen2.5-72B | 76.85 |
| 8 | Qwen 2 72B | 75.93 |
| 9 | Qwen 2.5 72B | 74.07 |
| 10 | Qwentile2.5-32B-Instruct | 74.07 |
| 11 | Gemma 3 27B (IT) | 73.15 |
| 12 | Qwen2.5-32B-Instruct-abliterated-v2 | 73.15 |
| 13 | Awqward2.5-32B-Instruct | 73.15 |
| 14 | Saka-14B | 73.15 |
| 15 | free-evo-qwen72B-v0.8-re | 73.15 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-jurisprudence · How the rankings work · Data refreshed daily, snapshot 2026-07-22.