Open Persian LLM - Part Multiple Choice — leaderboard

Metric: Accuracy (%). Source: huggingface.co. 31 models tracked.

Top models

#ModelScore
1Llama 3.3 70B Instruct51.42
2Gemma 3 27B (IT)48.56
3QwQ 32B-Preview46.64
4Qwen 2.5 32B Instruct46.06
5Gemma 2 27B (IT)46.03
6Gemma 3 12B (IT)43.52
7aya-expanse-32B43.36
8Gemma 2 9B (IT)42.7
9aya-23-35B36.79
10Qwen 2.5 7B Instruct36.72
11Llama 3.1 8B Instruct36.68
12Qwen 2 7B Instruct35.9
13Llama 3.1 8B35.52
14Hermes-3-Llama-3.1-8B35.01
15Llama 3 8B Instruct34.99

Interactive version: theaggregate.ai/benchmark?slug=open-persian-llm-part-multiple-choice · How the rankings work · Data refreshed daily, snapshot 2026-07-22.