Open Arabic LLM - Arabic MMLU Social Science (Middle School) — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | gemma-3-27B-pt | 60.58 |
| 2 | Saka-14B | 60.17 |
| 3 | Rombos-LLM-V2.5-Qwen-72B | 59.34 |
| 4 | Ultiima-72B | 58.92 |
| 5 | Qwen 2.5 14B | 58.51 |
| 6 | lambda-qwen2.5-14B-dpo-test | 58.09 |
| 7 | tempmotacilla-cinerea-0308 | 58.09 |
| 8 | Qwen 2.5 72B Instruct | 57.68 |
| 9 | Qwen 2.5 32B Instruct | 57.68 |
| 10 | free-evo-qwen72B-v0.8-re | 57.68 |
| 11 | Llama 3.3 70B Instruct | 57.26 |
| 12 | calme-2.1-qwen2.5-72B | 57.26 |
| 13 | calme-2.2-qwen2.5-72B | 57.26 |
| 14 | Awqward2.5-32B-Instruct | 56.85 |
| 15 | Qwen2.5-32B-Instruct-CFT | 56.85 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-social-science-middle-school · How the rankings work · Data refreshed daily, snapshot 2026-07-22.