Open Arabic LLM - Arabic MMLU HT Global Facts: leaderboard
The "Arabic MMLU HT Global Facts" 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 July 2028. 163 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2 72B | 54 | #521 |
| 2 | free-evo-qwen72B-v0.8-re | 54 | #435 |
| 3 | Saka-14B | 53 | #344 |
| 4 | Qwen 2.5 72B | 51 | #499 |
| 5 | Gemma 3 27B (IT) | 50 | #523 |
| 6 | Qwen 2.5 72B Instruct | 50 | #454 |
| 7 | gemma-3-27B-pt | 50 | #502 |
| 8 | AceGPT-v2-70B-Chat | 49 | #407 |
| 9 | Llama 3.3 70B Instruct | 47 | #534 |
| 10 | tempmotacilla-cinerea-0308 | 47 | #395 |
| 11 | Qwen 2.5 14B | 45 | #725 |
| 12 | Qwentile2.5-32B-Instruct | 44 | #339 |
| 13 | Hala-9B | 44 | #504 |
| 14 | Gemma 2 27B (IT) | 43 | #645 |
| 15 | AceGPT-v2-32B-Chat | 43 | #507 |
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-ht-global-facts · How It Works · Data refreshed daily, snapshot 2026-10-20.