Open Arabic LLM - Arabic MMLU HT Clinical Knowledge: leaderboard
The "Arabic MMLU HT Clinical Knowledge" 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 | 72.83 | #454 |
| 2 | AceGPT-v2-70B-Chat | 70.57 | #407 |
| 3 | Llama 3.3 70B Instruct | 68.68 | #534 |
| 4 | gemma-3-27B-pt | 68.68 | #502 |
| 5 | Qwentile2.5-32B-Instruct | 68.3 | #339 |
| 6 | Qwen 2.5 32B Instruct | 67.92 | #498 |
| 7 | Awqward2.5-32B-Instruct | 67.92 | #371 |
| 8 | AIC-1 | 67.92 | #231 |
| 9 | Qwen2.5-32B-Instruct-CFT | 67.55 | #375 |
| 10 | Gemma 3 27B (IT) | 67.17 | #523 |
| 11 | Qwen 2.5 72B | 66.42 | #499 |
| 12 | Qwen2.5-32B-Instruct-abliterated-v2 | 66.04 | #394 |
| 13 | tempmotacilla-cinerea-0308 | 66.04 | #395 |
| 14 | Saka-14B | 65.66 | #344 |
| 15 | Fanar-1-9B | 64.91 | #778 |
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-ht-clinical-knowledge · How It Works · Data refreshed daily, snapshot 2026-10-20.