Open Arabic LLM - Arabic MMLU General Knowledge (Primary School): leaderboard
The "Arabic MMLU General Knowledge (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 | Qwen 2.5 72B Instruct | 79.63 | #454 |
| 2 | gemma-3-27B-pt | 77.78 | #502 |
| 3 | Qwen 2.5 14B | 77.78 | #725 |
| 4 | gemma-3-12B-pt | 77.16 | #686 |
| 5 | AceGPT-v2-32B-Chat | 77.16 | #507 |
| 6 | Llama 3.3 70B Instruct | 75.93 | #534 |
| 7 | Qwen2.5-32B-Instruct-abliterated-v2 | 75.31 | #394 |
| 8 | Saka-14B | 74.07 | #344 |
| 9 | Qwen 2.5 32B Instruct | 73.46 | #498 |
| 10 | aya-23-35B | 73.46 | #883 |
| 11 | jais-adapted-70B-chat | 73.46 | #636 |
| 12 | Qwentile2.5-32B-Instruct | 73.46 | #339 |
| 13 | Fanar-1-9B-Instruct | 72.84 | #633 |
| 14 | EZO-Qwen2.5-32B-Instruct | 72.84 | #471 |
| 15 | Awqward2.5-32B-Instruct | 72.84 | #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-general-knowledge-primary-school · How It Works · Data refreshed daily, snapshot 2026-10-20.