Open Arabic LLM - Arabic MMLU HT Elementary Mathematics — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Rombos-LLM-V2.5-Qwen-72B | 71.69 |
| 2 | Ultiima-72B | 71.16 |
| 3 | lambda-qwen2.5-32B-dpo-test | 69.84 |
| 4 | Qwen2.5-32B-Instruct-abliterated-v2 | 67.46 |
| 5 | Saka-14B | 67.46 |
| 6 | calme-2.1-qwen2.5-72B | 66.67 |
| 7 | calme-2.2-qwen2.5-72B | 66.67 |
| 8 | Gemma 3 27B (IT) | 66.14 |
| 9 | Qwen3-8B-Base | 66.14 |
| 10 | Awqward2.5-32B-Instruct | 65.34 |
| 11 | tempmotacilla-cinerea-0308 | 64.81 |
| 12 | Qwen 2.5 32B Instruct | 64.55 |
| 13 | Qwen2.5-32B-Instruct-CFT | 64.29 |
| 14 | Qwen 2.5 72B Instruct | 64.02 |
| 15 | Qwentile2.5-32B-Instruct | 63.76 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-elementary-mathematics · How the rankings work · Data refreshed daily, snapshot 2026-07-22.