Open Arabic LLM - Arabic MMLU HT Logical Fallacies: leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 76.69 |
| 2 | Qwen 2 72B | 73.01 |
| 3 | Gemma 3 27B (IT) | 71.17 |
| 4 | Qwentile2.5-32B-Instruct | 71.17 |
| 5 | gemma-3-27B-pt | 70.55 |
| 6 | Awqward2.5-32B-Instruct | 69.33 |
| 7 | Llama 3.3 70B Instruct | 68.71 |
| 8 | Qwen 2.5 72B | 68.71 |
| 9 | Qwen2.5-32B-Instruct-abliterated-v2 | 68.71 |
| 10 | Gemma 2 27B (IT) | 68.1 |
| 11 | Qwen2.5-32B-Instruct-CFT | 67.48 |
| 12 | Qwen 2.5 32B Instruct | 66.87 |
| 13 | Saka-14B | 66.26 |
| 14 | EZO-Qwen2.5-32B-Instruct | 61.35 |
| 15 | aya-expanse-32B | 60.74 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-logical-fallacies · How It Works · Data refreshed daily, snapshot 2026-09-05.