Open Arabic LLM - Aratrust PhysicalHealth: leaderboard
The "Aratrust PhysicalHealth" 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. 162 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 72B Instruct | 94.52 | #454 |
| 2 | Llama 3.3 70B Instruct | 91.78 | #534 |
| 3 | Fanar-1-9B-Instruct | 91.78 | #633 |
| 4 | Qwen2.5-32B-Instruct-abliterated-v2 | 91.78 | #394 |
| 5 | AIC-1 | 91.78 | #231 |
| 6 | Gemma 3 27B (IT) | 90.41 | #523 |
| 7 | Qwen 2.5 72B | 90.41 | #499 |
| 8 | AceGPT-v2-32B-Chat | 90.41 | #507 |
| 9 | jais-adapted-70B-chat | 90.41 | #636 |
| 10 | AceGPT-v2-70B-Chat | 90.41 | #407 |
| 11 | Qwen 2 72B | 90.41 | #521 |
| 12 | Qwentile2.5-32B-Instruct | 90.41 | #339 |
| 13 | EZO-Qwen2.5-32B-Instruct | 90.41 | #471 |
| 14 | Awqward2.5-32B-Instruct | 90.41 | #371 |
| 15 | Qwen2.5-32B-Instruct-CFT | 90.41 | #375 |
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-aratrust-physicalhealth · How It Works · Data refreshed daily, snapshot 2026-10-20.