Open Arabic LLM - Arabic MMLU HT Prehistory — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Ultiima-72B | 81.17 |
| 2 | Rombos-LLM-V2.5-Qwen-72B | 80.56 |
| 3 | calme-2.1-qwen2.5-72B | 78.7 |
| 4 | calme-2.2-qwen2.5-72B | 78.09 |
| 5 | Qwen 2.5 72B Instruct | 75.93 |
| 6 | Llama 3.3 70B Instruct | 72.84 |
| 7 | Qwen 2.5 72B | 72.53 |
| 8 | Qwentile2.5-32B-Instruct | 70.68 |
| 9 | Awqward2.5-32B-Instruct | 69.14 |
| 10 | Gemma 3 27B (IT) | 68.52 |
| 11 | Qwen2.5-32B-Instruct-CFT | 68.21 |
| 12 | Qwen 2 72B | 68.21 |
| 13 | gemma-3-27B-pt | 67.9 |
| 14 | Qwen 2.5 32B Instruct | 67.9 |
| 15 | Qwen2.5-32B-Instruct-abliterated-v2 | 67.28 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-prehistory · How the rankings work · Data refreshed daily, snapshot 2026-07-22.