Open Arabic LLM - Arabic MMLU HT Virology: leaderboard
The "Arabic MMLU HT Virology" 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 2028. 163 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | gemma-3-27B-pt | 53.61 | #502 |
| 2 | Qwen 2.5 72B Instruct | 53.01 | #454 |
| 3 | Qwen 2.5 14B | 53.01 | #725 |
| 4 | Qwen2.5-32B-Instruct-abliterated-v2 | 53.01 | #394 |
| 5 | Qwen 2 72B | 52.41 | #521 |
| 6 | Qwen 2.5 32B Instruct | 51.2 | #498 |
| 7 | Qwen 2.5 32B | 51.2 | #546 |
| 8 | jais-adapted-70B-chat | 51.2 | #636 |
| 9 | Awqward2.5-32B-Instruct | 51.2 | #371 |
| 10 | Qwen2.5-32B-Instruct-CFT | 51.2 | #375 |
| 11 | free-evo-qwen72B-v0.8-re | 51.2 | #435 |
| 12 | Qwentile2.5-32B-Instruct | 50.6 | #339 |
| 13 | EZO-Qwen2.5-32B-Instruct | 50.6 | #471 |
| 14 | AIC-1 | 50.6 | #231 |
| 15 | emirati-14B-v2 | 50.6 | #726 |
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-virology · How It Works · Data refreshed daily, snapshot 2026-10-20.