Open Arabic LLM - Arabic MMLU Average — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Rombos-LLM-V2.5-Qwen-72B | 75.05 |
| 2 | Ultiima-72B | 75.04 |
| 3 | Qwen 2.5 72B Instruct | 71.43 |
| 4 | calme-2.1-qwen2.5-72B | 70.85 |
| 5 | Saka-14B | 70.84 |
| 6 | gemma-3-27B-pt | 70.76 |
| 7 | calme-2.2-qwen2.5-72B | 70.6 |
| 8 | lambda-qwen2.5-32B-dpo-test | 69.92 |
| 9 | Llama 3.3 70B Instruct | 69.85 |
| 10 | Qwen2.5-32B-Instruct-abliterated-v2 | 69.58 |
| 11 | Qwen 2.5 14B | 68.49 |
| 12 | Qwentile2.5-32B-Instruct | 68.24 |
| 13 | Awqward2.5-32B-Instruct | 68.08 |
| 14 | tempmotacilla-cinerea-0308 | 68.07 |
| 15 | Qwen 2.5 32B Instruct | 68.03 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-average · How the rankings work · Data refreshed daily, snapshot 2026-07-22.