Open Arabic LLM - Arabic MMLU Philosophy (High School): leaderboard
The "Arabic MMLU Philosophy (High School)" 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 | Llama 3.3 70B Instruct | 71.79 | #534 |
| 2 | Qwen 2.5 14B | 71.79 | #725 |
| 3 | tempmotacilla-cinerea-0308 | 71.79 | #395 |
| 4 | Qwen 2.5 72B Instruct | 69.23 | #454 |
| 5 | gemma-3-12B-pt | 69.23 | #686 |
| 6 | gemma-3-27B-pt | 69.23 | #502 |
| 7 | Qwen3-8B-Base | 69.23 | #626 |
| 8 | jais-adapted-70B-chat | 69.23 | #636 |
| 9 | Qwen3-4B-Base | 69.23 | #897 |
| 10 | lambda-qwen2.5-14B-dpo-test | 69.23 | #481 |
| 11 | Qwen2.5-32B-Instruct-abliterated-v2 | 69.23 | #394 |
| 12 | EZO-Qwen2.5-32B-Instruct | 69.23 | #471 |
| 13 | Saka-14B | 69.23 | #344 |
| 14 | free-evo-qwen72B-v0.8-re | 69.23 | #435 |
| 15 | AIC-1 | 69.23 | #231 |
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-philosophy-high-school · How It Works · Data refreshed daily, snapshot 2026-10-20.