Open Arabic LLM - Arabic MMLU HT Professional Psychology — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Ultiima-72B | 81.53 |
| 2 | Rombos-LLM-V2.5-Qwen-72B | 81.47 |
| 3 | calme-2.1-qwen2.5-72B | 81.13 |
| 4 | calme-2.2-qwen2.5-72B | 80.99 |
| 5 | Qwen 2.5 72B Instruct | 80.19 |
| 6 | Qwen 2 72B | 77.11 |
| 7 | Llama 3.3 70B Instruct | 74.94 |
| 8 | Qwen 2.5 72B | 74.79 |
| 9 | gemma-3-27B-pt | 73.82 |
| 10 | Gemma 3 27B (IT) | 73.25 |
| 11 | lambda-qwen2.5-32B-dpo-test | 72.37 |
| 12 | Awqward2.5-32B-Instruct | 72.05 |
| 13 | Qwen 2.5 32B Instruct | 71.94 |
| 14 | Qwentile2.5-32B-Instruct | 71.88 |
| 15 | Qwen2.5-32B-Instruct-CFT | 71.85 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-professional-psychology · How the rankings work · Data refreshed daily, snapshot 2026-07-22.