Open Arabic LLM - Arabic MMLU HT Astronomy — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | calme-2.2-qwen2.5-72B | 85.53 |
| 2 | calme-2.1-qwen2.5-72B | 84.87 |
| 3 | Rombos-LLM-V2.5-Qwen-72B | 84.21 |
| 4 | Qwen 2.5 72B Instruct | 83.55 |
| 5 | Ultiima-72B | 83.55 |
| 6 | Qwen 2.5 72B | 80.92 |
| 7 | Qwen 2.5 32B Instruct | 78.95 |
| 8 | Qwen2.5-32B-Instruct-abliterated-v2 | 78.95 |
| 9 | lambda-qwen2.5-32B-dpo-test | 78.95 |
| 10 | Llama 3.3 70B Instruct | 78.29 |
| 11 | Awqward2.5-32B-Instruct | 78.29 |
| 12 | tempmotacilla-cinerea-0308 | 78.29 |
| 13 | Qwentile2.5-32B-Instruct | 78.29 |
| 14 | Qwen2.5-32B-Instruct-CFT | 78.29 |
| 15 | Gemma 3 27B (IT) | 77.63 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-astronomy · How the rankings work · Data refreshed daily, snapshot 2026-07-22.