Open Arabic LLM - Arabic MMLU Geography (Primary School): leaderboard
The "Arabic MMLU Geography (Primary 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: Estimated already saturated. 163 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | AIC-1 | 82.46 | #231 |
| 2 | Qwen 2.5 72B Instruct | 80.7 | #454 |
| 3 | Qwen3-8B-Base | 78.95 | #626 |
| 4 | Qwentile2.5-32B-Instruct | 78.95 | #339 |
| 5 | Hala-9B | 78.95 | #504 |
| 6 | ALLaM-7B-Instruct-preview | 77.19 | #718 |
| 7 | AceGPT-v2-32B-Chat | 77.19 | #507 |
| 8 | Saka-14B | 77.19 | #344 |
| 9 | Qwen 2.5 32B Instruct | 75.44 | #498 |
| 10 | gemma-3-12B-pt | 75.44 | #686 |
| 11 | Qwen 2.5 14B | 75.44 | #725 |
| 12 | Qwen3-4B-Base | 75.44 | #897 |
| 13 | Qwen2.5-32B-Instruct-abliterated-v2 | 75.44 | #394 |
| 14 | EZO-Qwen2.5-32B-Instruct | 75.44 | #471 |
| 15 | Awqward2.5-32B-Instruct | 75.44 | #371 |
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-geography-primary-school · How It Works · Data refreshed daily, snapshot 2026-10-20.