Open Arabic LLM - Alghafa Multiple Choice Facts Truefalse Balanced Task: leaderboard
The "Alghafa Multiple Choice Facts Truefalse Balanced Task" 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 | aya-expanse-32B | 98.67 | #692 |
| 2 | Qwen 2 72B | 96 | #521 |
| 3 | c4ai-command-r7B-arabic-02-2025 | 96 | #732 |
| 4 | Hala-9B | 96 | #504 |
| 5 | Qwen 3 8B | 94.67 | #674 |
| 6 | ALLaM-7B-Instruct-preview | 94.67 | #718 |
| 7 | AceGPT-v2-32B-Chat | 94.67 | #507 |
| 8 | Fanar-1-9B-Instruct | 94.67 | #633 |
| 9 | Qwen2.5-32B-Instruct-abliterated-v2 | 94.67 | #394 |
| 10 | Yehia-7B-preview | 94.67 | #646 |
| 11 | AIC-1 | 94.67 | #231 |
| 12 | Qwen3-8B-Base | 93.33 | #626 |
| 13 | AceGPT-v2-8B-Chat | 93.33 | #736 |
| 14 | recoilme-gemma-2-9B-v0.2 | 93.33 | #629 |
| 15 | Gemma 3 27B (IT) | 92 | #523 |
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-alghafa-multiple-choice-facts-truefalse-balanced-task · How It Works · Data refreshed daily, snapshot 2026-10-20.