Open Arabic LLM - Arabic MMLU Computer Science (Middle School): leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen3-8B-Base | 96.3 |
| 2 | Qwen 2.5 14B | 96.3 |
| 3 | lambda-qwen2.5-14B-dpo-test | 96.3 |
| 4 | Saka-14B | 96.3 |
| 5 | Qwen 2 7B Instruct | 92.59 |
| 6 | Qwen 2.5 14B Instruct | 92.59 |
| 7 | Qwen2.5-32B-Instruct-abliterated-v2 | 92.59 |
| 8 | tempmotacilla-cinerea-0308 | 92.59 |
| 9 | Qwentile2.5-32B-Instruct | 92.59 |
| 10 | Qwen 2 72B | 92.59 |
| 11 | Qwen2.5-Lumen-14B | 92.59 |
| 12 | Qwen2.5-14B-Gutenberg-1e-Delta | 92.59 |
| 13 | Llama 3.3 70B Instruct | 88.89 |
| 14 | gemma-3-12B-pt | 88.89 |
| 15 | Qwen 2.5 32B Instruct | 88.89 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-computer-science-middle-school · How It Works · Data refreshed daily, snapshot 2026-09-05.