Open Persian LLM - MMLU Pro: leaderboard
The "MMLU Pro" subtask of Open Persian LLM. A task from the Open Persian LLM Leaderboard, evaluating models on Persian-language knowledge, reasoning and understanding.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 31 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Llama 3.3 70B Instruct | 43.67 | #534 |
| 2 | Gemma 3 27B (IT) | 40.1 | #523 |
| 3 | Qwen 2.5 32B Instruct | 38.19 | #498 |
| 4 | QwQ 32B-Preview | 37.94 | #651 |
| 5 | Gemma 2 27B (IT) | 36.28 | #645 |
| 6 | Gemma 2 9B (IT) | 31.74 | #788 |
| 7 | aya-expanse-32B | 31.03 | #692 |
| 8 | Gemma 3 12B (IT) | 29.36 | #664 |
| 9 | Llama 3 8B Instruct | 25.54 | #1129 |
| 10 | aya-23-35B | 25.54 | #883 |
| 11 | aya-expanse-8B | 25.06 | #931 |
| 12 | Llama 3.1 8B | 24.1 | #1151 |
| 13 | Qwen 2 7B Instruct | 23.87 | #991 |
| 14 | Qwen 2.5 7B Instruct | 21.96 | #862 |
| 15 | suzume-llama-3-8B-multilingual | 21.95 | #1084 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=open-persian-llm-mmlu-pro · How It Works · Data refreshed daily, snapshot 2026-10-20.