Open Arabic LLM - Arabic MMLU HT High School Geography: leaderboard
The "Arabic MMLU HT High School Geography" 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 | 86.36 | #454 |
| 2 | AceGPT-v2-70B-Chat | 84.34 | #407 |
| 3 | Llama 3.3 70B Instruct | 81.82 | #534 |
| 4 | AIC-1 | 81.82 | #231 |
| 5 | Qwen 2 72B | 81.31 | #521 |
| 6 | Qwen2.5-32B-Instruct-abliterated-v2 | 80.3 | #394 |
| 7 | Qwentile2.5-32B-Instruct | 80.3 | #339 |
| 8 | free-evo-qwen72B-v0.8-re | 79.8 | #435 |
| 9 | Saka-14B | 79.29 | #344 |
| 10 | Qwen 2.5 72B | 78.79 | #499 |
| 11 | AceGPT-v2-32B-Chat | 77.78 | #507 |
| 12 | Awqward2.5-32B-Instruct | 77.27 | #371 |
| 13 | tempmotacilla-cinerea-0308 | 77.27 | #395 |
| 14 | Rombos-LLM-V2.6-Qwen-14B | 77.27 | #516 |
| 15 | Qwen 2.5 32B Instruct | 76.77 | #498 |
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-geography · How It Works · Data refreshed daily, snapshot 2026-10-20.