Open Arabic LLM - Arabic MMLU Computer Science (Primary School): leaderboard
The "Arabic MMLU Computer Science (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 | Gemma 3 27B (IT) | 81.05 | #523 |
| 2 | gemma-3-27B-pt | 80.53 | #502 |
| 3 | AIC-1 | 80.53 | #231 |
| 4 | free-evo-qwen72B-v0.8-re | 80 | #435 |
| 5 | Qwen 2.5 7B | 79.47 | #922 |
| 6 | Qwen 2.5 72B | 79.47 | #499 |
| 7 | Qwen 2 72B | 78.95 | #521 |
| 8 | Qwen 2.5 72B Instruct | 77.89 | #454 |
| 9 | Qwen 2.5 14B | 77.37 | #725 |
| 10 | Saka-14B | 77.37 | #344 |
| 11 | aya-23-35B | 76.84 | #883 |
| 12 | Qwen3-8B-Base | 76.84 | #626 |
| 13 | Qwentile2.5-32B-Instruct | 76.84 | #339 |
| 14 | AceGPT-v2-32B | 76.84 | #768 |
| 15 | aya-expanse-32B | 76.32 | #692 |
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-computer-science-primary-school · How It Works · Data refreshed daily, snapshot 2026-10-20.