Open Arabic LLM - Madinah QA Average: leaderboard
The "Madinah QA Average" 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 January 2027. 163 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | AIC-1 | 78 | #231 |
| 2 | free-evo-qwen72B-v0.8-re | 74.29 | #435 |
| 3 | AceGPT-v2-32B-Chat | 73.77 | #507 |
| 4 | Qwentile2.5-32B-Instruct | 73.6 | #339 |
| 5 | Fanar-1-9B-Instruct | 73.35 | #633 |
| 6 | tempmotacilla-cinerea-0308 | 71.27 | #395 |
| 7 | Hala-9B | 71.23 | #504 |
| 8 | Saka-14B | 70.01 | #344 |
| 9 | Gemma 3 27B (IT) | 69.43 | #523 |
| 10 | recoilme-gemma-2-9B-v0.4 | 67.43 | #553 |
| 11 | jais-family-30B-16k-chat | 66.26 | #755 |
| 12 | AceGPT-v2-70B-Chat | 65.64 | #407 |
| 13 | Qwen 2.5 72B | 65.02 | #499 |
| 14 | Qwen2.5-7B-Instruct-Uncensored | 64.23 | #684 |
| 15 | c4ai-command-r7B-arabic-02-2025 | 63.84 | #732 |
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-madinah-qa-average · How It Works · Data refreshed daily, snapshot 2026-10-20.