Open Arabic LLM - Arabic MMLU HT Average: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 72.23 |
| 2 | Qwen 2 72B | 68.06 |
| 3 | Llama 3.3 70B Instruct | 67.79 |
| 4 | Gemma 3 27B (IT) | 66.92 |
| 5 | Qwentile2.5-32B-Instruct | 66.67 |
| 6 | gemma-3-27B-pt | 66.62 |
| 7 | Qwen 2.5 72B | 66.44 |
| 8 | Awqward2.5-32B-Instruct | 66.38 |
| 9 | Qwen 2.5 32B Instruct | 65.74 |
| 10 | Qwen2.5-32B-Instruct-abliterated-v2 | 65.73 |
| 11 | Qwen2.5-32B-Instruct-CFT | 65.73 |
| 12 | Saka-14B | 64.56 |
| 13 | free-evo-qwen72B-v0.8-re | 62.47 |
| 14 | tempmotacilla-cinerea-0308 | 62.45 |
| 15 | Qwen 2.5 14B | 61.69 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-average · How It Works · Data refreshed daily, snapshot 2026-09-05.