Open Arabic LLM - Arabic MMLU HT Virology — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | gemma-3-27B-pt | 53.61 |
| 2 | Qwen 2.5 72B Instruct | 53.01 |
| 3 | Qwen 2.5 14B | 53.01 |
| 4 | Qwen2.5-32B-Instruct-abliterated-v2 | 53.01 |
| 5 | Qwen 2 72B | 52.41 |
| 6 | lambda-qwen2.5-32B-dpo-test | 52.41 |
| 7 | Qwen 2.5 32B Instruct | 51.2 |
| 8 | Qwen 2.5 32B | 51.2 |
| 9 | Awqward2.5-32B-Instruct | 51.2 |
| 10 | Qwen2.5-32B-Instruct-CFT | 51.2 |
| 11 | free-evo-qwen72B-v0.8-re | 51.2 |
| 12 | Ultiima-72B | 51.2 |
| 13 | Qwentile2.5-32B-Instruct | 50.6 |
| 14 | EZO-Qwen2.5-32B-Instruct | 50.6 |
| 15 | calme-2.1-qwen2.5-72B | 50.6 |
Interactive version: theaggregate.ai/benchmark?slug=open-arabic-llm-arabic-mmlu-ht-virology · How the rankings work · Data refreshed daily, snapshot 2026-07-22.