Open Arabic LLM - Arabic MMLU HT Econometrics: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 57.02 |
| 2 | Llama 3.3 70B Instruct | 53.51 |
| 3 | Gemma 2 27B (IT) | 53.51 |
| 4 | Saka-14B | 53.51 |
| 5 | Qwen 2 72B | 51.75 |
| 6 | Gemma 3 27B (IT) | 50.88 |
| 7 | Qwen 2.5 72B | 50.88 |
| 8 | Qwen3-8B-Base | 48.25 |
| 9 | Qwen2.5-32B-Instruct-abliterated-v2 | 47.37 |
| 10 | tempmotacilla-cinerea-0308 | 47.37 |
| 11 | Qwen 2.5 32B Instruct | 46.49 |
| 12 | Qwen 2.5 14B | 46.49 |
| 13 | Awqward2.5-32B-Instruct | 46.49 |
| 14 | Qwen2.5-32B-Instruct-CFT | 46.49 |
| 15 | gemma-3-27B-pt | 44.74 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-econometrics · How It Works · Data refreshed daily, snapshot 2026-09-05.