Open Arabic LLM - Alghafa Multiple Choice Sentiment Task: leaderboard
The "Alghafa Multiple Choice Sentiment 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: Around 2033. 163 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 72B | 43.66 | #499 |
| 2 | jais-adapted-70B-chat | 43.6 | #636 |
| 3 | Qwen 2.5 72B Instruct | 43.02 | #454 |
| 4 | Qwen 2 72B | 43.02 | #521 |
| 5 | Saka-14B | 43.02 | #344 |
| 6 | SeaLLMs-v3-7B-Chat | 42.44 | #915 |
| 7 | Gemma 3 27B (IT) | 42.33 | #523 |
| 8 | Qwen 2.5 14B | 42.27 | #725 |
| 9 | Yehia-7B-preview | 41.69 | #646 |
| 10 | Qwen 2.5 32B | 41.63 | #546 |
| 11 | aya-expanse-32B | 41.57 | #692 |
| 12 | Hala-9B | 41.57 | #504 |
| 13 | Llama 3.1 8B Instruct | 41.51 | #1035 |
| 14 | Gemma 2 27B (IT) | 41.51 | #645 |
| 15 | ALLaM-7B-Instruct-preview | 41.4 | #718 |
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-sentiment-task · How It Works · Data refreshed daily, snapshot 2026-10-20.