Open Arabic LLM - Arabic MMLU Islamic Studies (Middle School): leaderboard
The "Arabic MMLU Islamic Studies (Middle School)" subtask of Open Arabic LLM. A task from the Open Arabic LLM Leaderboard, evaluating models on Arabic-language knowledge, reasoning and understanding across native and translated benchmarks.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around January 2027. 163 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 76.47 | #454 |
| 2 | Yehia-7B-preview | 76.47 | #646 |
| 3 | Llama 3.3 70B Instruct | 75.21 | #534 |
| 4 | gemma-3-27B-pt | 75.21 | #502 |
| 5 | ALLaM-7B-Instruct-preview | 75.21 | #718 |
| 6 | AceGPT-v2-32B-Chat | 73.11 | #507 |
| 7 | Saka-14B | 73.11 | #344 |
| 8 | aya-23-35B | 71.43 | #883 |
| 9 | jais-adapted-70B-chat | 71.43 | #636 |
| 10 | Awqward2.5-32B-Instruct | 71.43 | #371 |
| 11 | AIC-1 | 71.01 | #231 |
| 12 | lambda-qwen2.5-14B-dpo-test | 70.59 | #481 |
| 13 | Qwen2.5-32B-Instruct-CFT | 70.59 | #375 |
| 14 | Qwen 2.5 32B Instruct | 70.17 | #498 |
| 15 | EZO-Qwen2.5-32B-Instruct | 70.17 | #471 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-islamic-studies-middle-school · How It Works · Data refreshed daily, snapshot 2026-10-20.