Open Arabic LLM - Arabic MMLU History (High School): leaderboard
The "Arabic MMLU History (High School)" 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 September 2028. 163 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | AceGPT-v2-32B-Chat | 53.55 | #507 |
| 2 | Qwen 2.5 14B | 53.42 | #725 |
| 3 | gemma-3-27B-pt | 53.16 | #502 |
| 4 | Saka-14B | 52.24 | #344 |
| 5 | AceGPT-v2-70B-Chat | 52.11 | #407 |
| 6 | AIC-1 | 52.11 | #231 |
| 7 | Llama 3.3 70B Instruct | 51.84 | #534 |
| 8 | Qwen 2.5 72B Instruct | 51.71 | #454 |
| 9 | tempmotacilla-cinerea-0308 | 51.58 | #395 |
| 10 | free-evo-qwen72B-v0.8-re | 50.79 | #435 |
| 11 | AceGPT-v2-32B | 49.61 | #768 |
| 12 | AceGPT-v2-8B-Chat | 49.21 | #736 |
| 13 | gemma-3-12B-pt | 48.68 | #686 |
| 14 | Qwen 2.5 72B | 48.55 | #499 |
| 15 | Fanar-1-9B | 47.89 | #778 |
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-history-high-school · How It Works · Data refreshed daily, snapshot 2026-10-20.