Open Arabic LLM - Arabic MMLU HT Formal Logic Leaderboard

Open Arabic LLM - Arabic MMLU HT Formal Logic — leaderboard

Metric: Accuracy (%). Source: huggingface.co. 163 models tracked.

#	Model	Score
1	Qwen3-8B-Base	61.9
2	Ultiima-72B	61.11
3	Qwen2.5-32B-Instruct-CFT	59.52
4	calme-2.1-qwen2.5-72B	59.52
5	calme-2.2-qwen2.5-72B	59.52
6	Rombos-LLM-V2.5-Qwen-72B	59.52
7	Gemma 3 27B (IT)	58.73
8	Llama 3.3 70B Instruct	58.73
9	lambda-qwen2.5-14B-dpo-test	58.73
10	Awqward2.5-32B-Instruct	57.94
11	tempmotacilla-cinerea-0308	57.94
12	lambda-qwen2.5-32B-dpo-test	57.94
13	Qwen 2.5 32B Instruct	57.14
14	Qwen2.5-32B-Instruct-abliterated-v2	57.14
15	Qwen 2.5 72B Instruct	56.35