Open Persian LLM Leaderboard
Models answer Persian multiple-choice questions. The leaderboard averages accuracy across five tasks, including two question sets from the organizers.
The project description also lists GSM Persian. The scoring code omits it, and the published results contain five task scores for each model.
- Made by
- Part AI and Amirkabir University of Technology NLP Lab
- Released
- November 2024
- Kind
- Knowledge in many languages
- Level
- University
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- Public, plus a private test set
- Measures
- Capability
- Language
- Persian
- Interaction
- Single turn
- Items from
- Another benchmark, Written by experts
- Refresh
- One fixed set
- Run by
- A host
- Strongest model tested
-
Gemma 3 27B #165 on the capability index
- Last model added
- March 2025
- Scores refreshed
- November 20th, 2026
Llama 3.3 70B Instruct leads with 70.4% among models released by March 2025 Llama 3.3 70B Instruct leads with 70.4% among models released by March 2025
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in March 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 70.4
Llama 3.3 70B Instruct
- 67.8
Gemma 3 27B
- 65.5
Gemma 2 27B
- 64.8
QwQ-32B-Preview
- 64.5
Qwen2.5-32B
- 62.9
Gemma 2 9B
- 61.9
Aya Expanse 32B
- 60.9
Gemma 3 12B
- 56.8 CohereForAI/aya-23-35B
- 53.7
Aya Expanse 8B
- 51.9
Qwen2.5-7B
- 51.4
Qwen2-7B-Instruct
- 50.7 PartAI/Dorna2-Llama3.1-8B-Instruct
- 50.2 O1-OPEN/OpenO1-LLama-8B-v0.1
- 50.1
Llama 3.1-8B
- 49.8 CohereForAI/aya-23-8B
- 49.1 PrincetonPLI/Llama-3-8B-Instruct-SkillMix
- 49.1
Gemma 3 4B
- 49.0
Llama 3-8B
- 48.8 NousResearch/Hermes-3-Llama-3.1-8B
- 48.3 PartAI/Dorna-Llama3-8B-Instruct
- 47.3 lightblue/suzume-llama-3-8B-multilingual
- 45.3
Gemma 2 2B IT
- 41.5 amirakhlaghiqqq/PersianPhi
- 39.5 universitytehran/PersianMind-v1.0
- 34.1 Skywork/Skywork-o1-Open-Llama-3.1-8B
- 32.7
Llama 3.2 1B
- 30.8 MaralGPT/Maral-7B-alpha-1
- 30.7
Gemma 3 1B
- 1
Llama 3.3 70B Instruct bfloat16 70.4%
- 2
Gemma 3 27B bfloat16 67.8%
- 3
Gemma 2 27B bfloat16 65.5%
- 4
QwQ-32B-Preview bfloat16 64.8%
- 5
Qwen2.5-32B bfloat16 64.5%
- 6
Gemma 2 9B bfloat16 62.9%
- 7
Aya Expanse 32B float16 61.9%
- 8
Gemma 3 12B bfloat16 60.9%
- 9 CohereForAI/aya-23-35B float16 56.8%
- 10
Aya Expanse 8B float16 53.7%
- 11
Qwen2.5-7B bfloat16 51.9%
- 12
Qwen2-7B-Instruct bfloat16 51.4%
- 13 PartAI/Dorna2-Llama3.1-8B-Instruct bfloat16 50.7%
- 14 O1-OPEN/OpenO1-LLama-8B-v0.1 bfloat16 50.2%
- 15
Llama 3.1-8B bfloat16 50.1%
- 16 CohereForAI/aya-23-8B float16 49.8%
- 17 PrincetonPLI/Llama-3-8B-Instruct-SkillMix bfloat16 49.1%
- 18
Gemma 3 4B bfloat16 49.1%
- 19
Llama 3-8B bfloat16 49%
- 20 NousResearch/Hermes-3-Llama-3.1-8B bfloat16 48.8%
- 21 PartAI/Dorna-Llama3-8B-Instruct bfloat16 48.3%
- 22 lightblue/suzume-llama-3-8B-multilingual bfloat16 47.3%
- 23
Gemma 2 2B IT bfloat16 45.3%
- 24 amirakhlaghiqqq/PersianPhi bfloat16 41.5%
- 25 universitytehran/PersianMind-v1.0 bfloat16 39.5%
- 26 Skywork/Skywork-o1-Open-Llama-3.1-8B bfloat16 34.1%
- 27
Llama 3.2 1B bfloat16 32.7%
- 28 MaralGPT/Maral-7B-alpha-1 bfloat16 30.8%
- 29
Gemma 3 1B bfloat16 30.7%
Source: Official Open Persian LLM Leaderboard leaderboard
- 70.4
Llama 3.3 70B Instruct
- 67.8
Gemma 3 27B
- 65.5
Gemma 2 27B
- 64.8
QwQ-32B-Preview
- 64.5
Qwen2.5-32B
- 62.9
Gemma 2 9B
- 61.9
Aya Expanse 32B
- 60.9
Gemma 3 12B
- 53.7
Aya Expanse 8B
- 51.9
Qwen2.5-7B
- 51.4
Qwen2-7B-Instruct
- 50.1
Llama 3.1-8B
- 49.1
Gemma 3 4B
- 49.0
Llama 3-8B
- 45.3
Gemma 2 2B IT
- 32.7
Llama 3.2 1B
- 30.7
Gemma 3 1B
- 1
Llama 3.3 70B Instruct bfloat16 70.4%
- 2
Gemma 3 27B bfloat16 67.8%
- 3
Gemma 2 27B bfloat16 65.5%
- 4
QwQ-32B-Preview bfloat16 64.8%
- 5
Qwen2.5-32B bfloat16 64.5%
- 6
Gemma 2 9B bfloat16 62.9%
- 7
Aya Expanse 32B float16 61.9%
- 8
Gemma 3 12B bfloat16 60.9%
- 9
Aya Expanse 8B float16 53.7%
- 10
Qwen2.5-7B bfloat16 51.9%
- 11
Qwen2-7B-Instruct bfloat16 51.4%
- 12
Llama 3.1-8B bfloat16 50.1%
- 13
Gemma 3 4B bfloat16 49.1%
- 14
Llama 3-8B bfloat16 49%
- 15
Gemma 2 2B IT bfloat16 45.3%
- 16
Llama 3.2 1B bfloat16 32.7%
- 17
Gemma 3 1B bfloat16 30.7%
Source: Official Open Persian LLM Leaderboard leaderboard
| # | Entry | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | | 70.4% | bfloat16 | Dec 2024 | huggingface.co | Official leaderboard |
| 2 | | 67.8% | bfloat16 | Mar 2025 | huggingface.co | Official leaderboard |
| 3 | | 65.5% | bfloat16 | Jun 2024 | huggingface.co | Official leaderboard |
| 4 | | 64.8% | bfloat16 | Nov 2024 | huggingface.co | Official leaderboard |
| 5 | | 64.5% | bfloat16 | Sep 2024 | huggingface.co | Official leaderboard |
| 6 | | 62.9% | bfloat16 | Jun 2024 | huggingface.co | Official leaderboard |
| 7 | | 61.9% | float16 | Oct 2024 | huggingface.co | Official leaderboard |
| 8 | | 60.9% | bfloat16 | Mar 2025 | huggingface.co | Official leaderboard |
| 9 | CohereForAI/aya-23-35B 35B | 56.8% | float16 | huggingface.co | Official leaderboard | |
| 10 | | 53.7% | float16 | Oct 2024 | huggingface.co | Official leaderboard |
| 11 | | 51.9% | bfloat16 | Sep 2024 | huggingface.co | Official leaderboard |
| 12 | | 51.4% | bfloat16 | Jun 2024 | huggingface.co | Official leaderboard |
| 13 | PartAI/Dorna2-Llama3.1-8B-Instruct 8B | 50.7% | bfloat16 | huggingface.co | Official leaderboard | |
| 14 | O1-OPEN/OpenO1-LLama-8B-v0.1 8B | 50.2% | bfloat16 | huggingface.co | Official leaderboard | |
| 15 | | 50.1% | bfloat16 | Jul 2024 | huggingface.co | Official leaderboard |
| 16 | CohereForAI/aya-23-8B 8B | 49.8% | float16 | huggingface.co | Official leaderboard | |
| 17 | PrincetonPLI/Llama-3-8B-Instruct-SkillMix 8B | 49.1% | bfloat16 | huggingface.co | Official leaderboard | |
| 18 | | 49.1% | bfloat16 | Mar 2025 | huggingface.co | Official leaderboard |
| 19 | | 49% | bfloat16 | Apr 2024 | huggingface.co | Official leaderboard |
| 20 | NousResearch/Hermes-3-Llama-3.1-8B 8B | 48.8% | bfloat16 | huggingface.co | Official leaderboard | |
| 21 | PartAI/Dorna-Llama3-8B-Instruct 8B | 48.3% | bfloat16 | huggingface.co | Official leaderboard | |
| 22 | lightblue/suzume-llama-3-8B-multilingual 8B | 47.3% | bfloat16 | huggingface.co | Official leaderboard | |
| 23 | | 45.3% | bfloat16 | Jul 2024 | huggingface.co | Official leaderboard |
| 24 | amirakhlaghiqqq/PersianPhi 3.9B | 41.5% | bfloat16 | huggingface.co | Official leaderboard | |
| 25 | universitytehran/PersianMind-v1.0 6.8B | 39.5% | bfloat16 | huggingface.co | Official leaderboard | |
| 26 | Skywork/Skywork-o1-Open-Llama-3.1-8B 8B | 34.1% | bfloat16 | huggingface.co | Official leaderboard | |
| 27 | | 32.7% | bfloat16 | Sep 2024 | huggingface.co | Official leaderboard |
| 28 | MaralGPT/Maral-7B-alpha-1 7.2B | 30.8% | bfloat16 | huggingface.co | Official leaderboard | |
| 29 | | 30.7% | bfloat16 | Mar 2025 | huggingface.co | Official leaderboard |
| # | Model | Score | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | | 70.4% | bfloat16 | Dec 2024 | huggingface.co | Official leaderboard |
| 2 | | 67.8% | bfloat16 | Mar 2025 | huggingface.co | Official leaderboard |
| 3 | | 65.5% | bfloat16 | Jun 2024 | huggingface.co | Official leaderboard |
| 4 | | 64.8% | bfloat16 | Nov 2024 | huggingface.co | Official leaderboard |
| 5 | | 64.5% | bfloat16 | Sep 2024 | huggingface.co | Official leaderboard |
| 6 | | 62.9% | bfloat16 | Jun 2024 | huggingface.co | Official leaderboard |
| 7 | | 61.9% | float16 | Oct 2024 | huggingface.co | Official leaderboard |
| 8 | | 60.9% | bfloat16 | Mar 2025 | huggingface.co | Official leaderboard |
| 9 | | 53.7% | float16 | Oct 2024 | huggingface.co | Official leaderboard |
| 10 | | 51.9% | bfloat16 | Sep 2024 | huggingface.co | Official leaderboard |
| 11 | | 51.4% | bfloat16 | Jun 2024 | huggingface.co | Official leaderboard |
| 12 | | 50.1% | bfloat16 | Jul 2024 | huggingface.co | Official leaderboard |
| 13 | | 49.1% | bfloat16 | Mar 2025 | huggingface.co | Official leaderboard |
| 14 | | 49% | bfloat16 | Apr 2024 | huggingface.co | Official leaderboard |
| 15 | | 45.3% | bfloat16 | Jul 2024 | huggingface.co | Official leaderboard |
| 16 | | 32.7% | bfloat16 | Sep 2024 | huggingface.co | Official leaderboard |
| 17 | | 30.7% | bfloat16 | Mar 2025 | huggingface.co | Official leaderboard |
Rankings by part
We read 5 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Part Multiple Choice, ARC Easy and ARC Challenge. Llama 3.3 70B Instruct leads 3 of them, and 2 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
Part Multiple Choice
17 models we track have a score here.
- 1
Llama 3.3 70B Instruct 51.4%
- 2
Gemma 3 27B 48.6%
- 3
QwQ-32B-Preview ↑1 46.6%
- 4
Qwen2.5-32B ↑1 46.1%
- 5
Gemma 2 27B ↓2 46%
- 6
Gemma 3 12B ↑2 43.5%
- 7
Aya Expanse 32B 43.4%
- 8
Gemma 2 9B ↓2 42.7%
- 9
Qwen2.5-7B ↑1 36.7%
- 10
Llama 3.1-8B ↑2 36.7%
ARC Easy
17 models we track have a score here.
- 1
Gemma 2 27B ↑2 96%
- 2
Gemma 3 27B 95.7%
- 3
Llama 3.3 70B Instruct ↓2 95.4%
- 4
Gemma 3 12B ↑4 93.4%
- 5
Gemma 2 9B ↑1 93.1%
- 6
Aya Expanse 32B ↑1 93.1%
- 7
QwQ-32B-Preview ↓3 92%
- 8
Qwen2.5-32B ↓3 90.8%
- 9
Aya Expanse 8B 79.6%
- 10
Qwen2.5-7B 79%
ARC Challenge
17 models we track have a score here.
- 1
Gemma 3 27B ↑1 90.6%
- 2
Llama 3.3 70B Instruct ↓1 89.9%
- 3
QwQ-32B-Preview ↑1 87.2%
- 4
Gemma 2 27B ↓1 85.9%
- 5
Qwen2.5-32B 85.9%
- 6
Gemma 2 9B 84.6%
- 7
Gemma 3 12B ↑1 81.2%
- 8
Aya Expanse 32B ↓1 79.9%
- 9
Aya Expanse 8B 70.5%
- 10
Qwen2.5-7B 69.1%
MMLU Pro
17 models we track have a score here.
- 1
Llama 3.3 70B Instruct 43.7%
- 2
Gemma 3 27B 40.1%
- 3
Qwen2.5-32B ↑2 38.2%
- 4
QwQ-32B-Preview 37.9%
- 5
Gemma 2 27B ↓2 36.3%
- 6
Gemma 2 9B 31.7%
- 7
Aya Expanse 32B 31%
- 8
Gemma 3 12B 29.4%
- 9
Llama 3-8B ↑5 25.5%
- 10
Aya Expanse 8B ↓1 25.1%
AUT Multiple Choice Persian
17 models we track have a score here.
- 1
Llama 3.3 70B Instruct 71.4%
- 2
Gemma 3 27B 64.3%
- 3
Gemma 2 27B 63.1%
- 4
Gemma 2 9B ↑2 62.3%
- 5
Aya Expanse 32B ↑2 62.3%
- 6
Qwen2.5-32B ↓1 61.3%
- 7
QwQ-32B-Preview ↓3 60.1%
- 8
Aya Expanse 8B ↑1 58.4%
- 9
Gemma 3 12B ↓1 57%
- 10
Llama 3.1-8B ↑2 54.2%
2 models have held number one on the official leaderboard since its release in November 2024
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Open Persian LLM Leaderboard leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Open Persian LLM Leaderboard leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Multilingual
- MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
- WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET.
K-EXAONE 90.5%
- MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language.
Qwen3.7-Max 87%
- CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories.
Solar Open 2 90.7%
- Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
- MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%
Data audit
9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 20 months ago, in March 2025, for Gemma 3 12B. Every benchmark
Catalog record 2 pass
- Pass Task access Public tasks, plus a private test set.
- Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
- Pass Independent runs Scores from the official leaderboard (29 entries).
- Pass Models scored 17 models we track, plus 12 other entries on its leaderboard.
- Warning Newest score The newest came 20 months ago, in March 2025, for Gemma 3 12B.
- Note Frontier models None of the top 10 has a score. The strongest tested is Gemma 3 27B (#165).
- Pass Headroom The best is 70.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on huggingface.co: 19 rows for models we track and 12 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.