Open Arabic LLM - Arabic MMLU HT High School Psychology: leaderboard
The "Arabic MMLU HT High School Psychology" 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 | 84.95 | #454 |
| 2 | AceGPT-v2-70B-Chat | 84.22 | #407 |
| 3 | Gemma 3 27B (IT) | 83.12 | #523 |
| 4 | Qwen 2 72B | 82.57 | #521 |
| 5 | gemma-3-27B-pt | 80.55 | #502 |
| 6 | Gemma 2 27B (IT) | 80.18 | #645 |
| 7 | Qwen 2.5 72B | 79.45 | #499 |
| 8 | Llama 3.3 70B Instruct | 79.08 | #534 |
| 9 | Hala-9B | 78.53 | #504 |
| 10 | Qwentile2.5-32B-Instruct | 78.35 | #339 |
| 11 | AIC-1 | 78.35 | #231 |
| 12 | Awqward2.5-32B-Instruct | 77.43 | #371 |
| 13 | Fanar-1-9B | 77.25 | #778 |
| 14 | Qwen 2.5 32B Instruct | 76.7 | #498 |
| 15 | gemma-3-12B-pt | 76.51 | #686 |
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-high-school-psychology · How It Works · Data refreshed daily, snapshot 2026-10-20.