Persian IFEval
Models respond to 457 Persian prompts while following constraints on wording, length, or structure.
MIZAN ranks 45 model entries by strict instruction accuracy. The release contains 681 individual instructions.
- Made by
- MCINext
- Released
- May 2025
- Kind
- Chat and instructions in many languages
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Language
- Persian
- Interaction
- Single turn
- Items from
- Another benchmark, Translated, Generated
- Refresh
- One fixed set
- Run by
- Anyone
- Based on
- IFEval
- Strongest model tested
- o3 #84 on the capability index
- Last model added
- September 2025
- Scores refreshed
- November 20th, 2026
o3 leads with 92.6% among models released by September 2025 o3 leads with 92.6% among models released by September 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 September 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
- 92.6 o3
- 91.0 GPT-5 mini
- 89.7 GPT-5 nano
- 89.1
Gemini 2.5 Pro
- 88.7 gpt-oss-120b
- 88.1 Claude 3.7 Sonnet
- 88.0
Gemini 2.5 Flash
- 86.3 GPT-4.1
- 85.9 gpt-oss-20b
- 85.0
Gemini 2.0 Flash
- 84.4 c4ai-command-a-03-2025
- 84.1
Gemini 2.0 Flash-Lite
- 83.7 deepseek-reasoner
- 83.4 GPT-4.1 mini
- 83.3
Qwen3-30B-A3B
- 83.1 deepseek-chat
- 83.0
Gemma 3 27B
- 83.0 GPT-4o
- 82.5 Claude 3.5 Haiku
- 81.5
Gemma 3 12B
- 81.1
Qwen3-14B
- 80.3
Qwen3-32B
- 80.2 GPT-4o mini
- 79.0
Llama 4 Scout
- 75.8 GPT-4.1 nano
- 75.8
Qwen3-4B
- 75.3
Mistral Small 3.1
- 75.0
Gemma 3n E4B
- 74.7
Qwen3-8B
- 74.5
Gemma 3 4B
- 71.3
Llama 3.3 70B Instruct
- 70.1
Command R+
- 69.9
Aya Expanse 32B
- 68.0 deepseek-r1-0528-qwen3-8b
- 60.4 yandexgpt-lite
- 57.9
Command R (03-2024)
- 54.5
Gemma 3 1B
- 53.3
Llama 3.2 3B
- 44.0 Mistral-7B-Instruct-v0.3
- 36.6
Llama 3.2 1B
- 30.8
Gemma 3 270M
- 1 o3 Thinking 92.6%
- 2 GPT-5 mini Thinking 91%
- 3 GPT-5 nano Thinking 89.7%
- 4
Gemini 2.5 Pro Jun 2025, Thinking 89.1%
- 5 gpt-oss-120b Thinking 88.7%
- 6 Claude 3.7 Sonnet Non-thinking 88.1%
- 7
Gemini 2.5 Flash Sep 2025, Thinking 88%
- 8 GPT-4.1 Non-thinking 86.3%
- 9 gpt-oss-20b Thinking 85.9%
- 10
Gemini 2.0 Flash Feb 2025, Non-thinking 85%
- 11 c4ai-command-a-03-2025 Non-thinking 84.4%
- 12
Gemini 2.0 Flash-Lite Non-thinking 84.1%
- 13 deepseek-reasoner Thinking 83.7%
- 14 GPT-4.1 mini Non-thinking 83.4%
- 15
Qwen3-30B-A3B Non-thinking 83.3%
- 16 deepseek-chat Non-thinking 83.1%
- 17
Gemma 3 27B Non-thinking 83%
- 18 GPT-4o Mar 2025, Non-thinking 83%
- 19 Claude 3.5 Haiku Non-thinking 82.5%
- 20
Gemma 3 12B Non-thinking 81.5%
- 21
Qwen3-14B Non-thinking 81.1%
- 22
Qwen3-32B Non-thinking 80.3%
- 23 GPT-4o mini Non-thinking 80.2%
- 24
Llama 4 Scout Non-thinking 79%
- 25 GPT-4.1 nano Non-thinking 75.8%
- 26
Qwen3-4B Non-thinking 75.8%
- 27
Mistral Small 3.1 Non-thinking 75.3%
- 28
Gemma 3n E4B Non-thinking 75%
- 29
Qwen3-8B Non-thinking 74.7%
- 30
Gemma 3 4B Non-thinking 74.5%
- 31
Llama 3.3 70B Instruct Non-thinking 71.3%
- 32
Command R+ Non-thinking 70.1%
- 33
Aya Expanse 32B Non-thinking 69.9%
- 34 deepseek-r1-0528-qwen3-8b Thinking 68%
- 35 yandexgpt-lite Non-thinking 60.4%
- 36
Command R (03-2024) Non-thinking 57.9%
- 37
Gemma 3 1B Non-thinking 54.5%
- 38
Llama 3.2 3B Non-thinking 53.3%
- 39 Mistral-7B-Instruct-v0.3 Non-thinking 44%
- 40
Llama 3.2 1B Non-thinking 36.6%
- 41
Gemma 3 270M Non-thinking 30.8%
Source: Official Persian IFEval leaderboard
- 92.6 o3
- 91.0 GPT-5 mini
- 89.7 GPT-5 nano
- 89.1
Gemini 2.5 Pro
- 88.7 gpt-oss-120b
- 88.1 Claude 3.7 Sonnet
- 88.0
Gemini 2.5 Flash
- 86.3 GPT-4.1
- 85.9 gpt-oss-20b
- 85.0
Gemini 2.0 Flash
- 84.1
Gemini 2.0 Flash-Lite
- 83.4 GPT-4.1 mini
- 83.3
Qwen3-30B-A3B
- 83.0
Gemma 3 27B
- 83.0 GPT-4o
- 82.5 Claude 3.5 Haiku
- 81.5
Gemma 3 12B
- 81.1
Qwen3-14B
- 80.3
Qwen3-32B
- 80.2 GPT-4o mini
- 79.0
Llama 4 Scout
- 75.8 GPT-4.1 nano
- 75.8
Qwen3-4B
- 75.3
Mistral Small 3.1
- 75.0
Gemma 3n E4B
- 74.7
Qwen3-8B
- 74.5
Gemma 3 4B
- 71.3
Llama 3.3 70B Instruct
- 70.1
Command R+
- 69.9
Aya Expanse 32B
- 68.0 deepseek-r1-0528-qwen3-8b
- 57.9
Command R (03-2024)
- 54.5
Gemma 3 1B
- 53.3
Llama 3.2 3B
- 36.6
Llama 3.2 1B
- 30.8
Gemma 3 270M
- 1 o3 Thinking 92.6%
- 2 GPT-5 mini Thinking 91%
- 3 GPT-5 nano Thinking 89.7%
- 4
Gemini 2.5 Pro Jun 2025, Thinking 89.1%
- 5 gpt-oss-120b Thinking 88.7%
- 6 Claude 3.7 Sonnet Non-thinking 88.1%
- 7
Gemini 2.5 Flash Sep 2025, Thinking 88%
- 8 GPT-4.1 Non-thinking 86.3%
- 9 gpt-oss-20b Thinking 85.9%
- 10
Gemini 2.0 Flash Feb 2025, Non-thinking 85%
- 11
Gemini 2.0 Flash-Lite Non-thinking 84.1%
- 12 GPT-4.1 mini Non-thinking 83.4%
- 13
Qwen3-30B-A3B Non-thinking 83.3%
- 14
Gemma 3 27B Non-thinking 83%
- 15 GPT-4o Mar 2025, Non-thinking 83%
- 16 Claude 3.5 Haiku Non-thinking 82.5%
- 17
Gemma 3 12B Non-thinking 81.5%
- 18
Qwen3-14B Non-thinking 81.1%
- 19
Qwen3-32B Non-thinking 80.3%
- 20 GPT-4o mini Non-thinking 80.2%
- 21
Llama 4 Scout Non-thinking 79%
- 22 GPT-4.1 nano Non-thinking 75.8%
- 23
Qwen3-4B Non-thinking 75.8%
- 24
Mistral Small 3.1 Non-thinking 75.3%
- 25
Gemma 3n E4B Non-thinking 75%
- 26
Qwen3-8B Non-thinking 74.7%
- 27
Gemma 3 4B Non-thinking 74.5%
- 28
Llama 3.3 70B Instruct Non-thinking 71.3%
- 29
Command R+ Non-thinking 70.1%
- 30
Aya Expanse 32B Non-thinking 69.9%
- 31 deepseek-r1-0528-qwen3-8b Thinking 68%
- 32
Command R (03-2024) Non-thinking 57.9%
- 33
Gemma 3 1B Non-thinking 54.5%
- 34
Llama 3.2 3B Non-thinking 53.3%
- 35
Llama 3.2 1B Non-thinking 36.6%
- 36
Gemma 3 270M Non-thinking 30.8%
Source: Official Persian IFEval leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | o3 | 92.6% | Thinking | Apr 2025 | Official leaderboard |
| 2 | GPT-5 mini | 91% | Thinking | Aug 2025 | Official leaderboard |
| 3 | GPT-5 nano | 89.7% | Thinking | Aug 2025 | Official leaderboard |
| 4 | | 89.1% | Thinking | Jun 2025 | Official leaderboard |
| 5 | gpt-oss-120b | 88.7% | Thinking | Aug 2025 | Official leaderboard |
| 6 | Claude 3.7 Sonnet | 88.1% | Non-thinking | Feb 2025 | Official leaderboard |
| 7 | | 88% | Thinking | Sep 2025 | Official leaderboard |
| 8 | GPT-4.1 | 86.3% | Non-thinking | Apr 2025 | Official leaderboard |
| 9 | gpt-oss-20b | 85.9% | Thinking | Aug 2025 | Official leaderboard |
| 10 | | 85% | Non-thinking | Feb 2025 | Official leaderboard |
| 11 | c4ai-command-a-03-2025 open weights, 111B | 84.4% | Non-thinking | Official leaderboard | |
| 12 | | 84.1% | Non-thinking | Feb 2025 | Official leaderboard |
| 13 | deepseek-reasoner open weights, 671B | 83.7% | Thinking | Official leaderboard | |
| 14 | GPT-4.1 mini | 83.4% | Non-thinking | Apr 2025 | Official leaderboard |
| 15 | | 83.3% | Non-thinking | Apr 2025 | Official leaderboard |
| 16 | deepseek-chat open weights, 671B | 83.1% | Non-thinking | Official leaderboard | |
| 17 | | 83% | Non-thinking | Mar 2025 | Official leaderboard |
| 18 | GPT-4o Mar 2025 | 83% | Non-thinking | Mar 2025 | Official leaderboard |
| 19 | Claude 3.5 Haiku | 82.5% | Non-thinking | Oct 2024 | Official leaderboard |
| 20 | | 81.5% | Non-thinking | Mar 2025 | Official leaderboard |
| 21 | | 81.1% | Non-thinking | Apr 2025 | Official leaderboard |
| 22 | | 80.3% | Non-thinking | Apr 2025 | Official leaderboard |
| 23 | GPT-4o mini | 80.2% | Non-thinking | Jul 2024 | Official leaderboard |
| 24 | | 79% | Non-thinking | Apr 2025 | Official leaderboard |
| 25 | GPT-4.1 nano | 75.8% | Non-thinking | Apr 2025 | Official leaderboard |
| 26 | | 75.8% | Non-thinking | Apr 2025 | Official leaderboard |
| 27 | | 75.3% | Non-thinking | Mar 2025 | Official leaderboard |
| 28 | | 75% | Non-thinking | Jun 2025 | Official leaderboard |
| 29 | | 74.7% | Non-thinking | Apr 2025 | Official leaderboard |
| 30 | | 74.5% | Non-thinking | Mar 2025 | Official leaderboard |
| 31 | | 71.3% | Non-thinking | Dec 2024 | Official leaderboard |
| 32 | | 70.1% | Non-thinking | Aug 2024 | Official leaderboard |
| 33 | | 69.9% | Non-thinking | Oct 2024 | Official leaderboard |
| 34 | deepseek-r1-0528-qwen3-8b | 68% | Thinking | May 2025 | Official leaderboard |
| 35 | yandexgpt-lite open weights, 8B | 60.4% | Non-thinking | Official leaderboard | |
| 36 | | 57.9% | Non-thinking | Mar 2024 | Official leaderboard |
| 37 | | 54.5% | Non-thinking | Mar 2025 | Official leaderboard |
| 38 | | 53.3% | Non-thinking | Sep 2024 | Official leaderboard |
| 39 | Mistral-7B-Instruct-v0.3 open weights, 7.3B | 44% | Non-thinking | Official leaderboard | |
| 40 | | 36.6% | Non-thinking | Sep 2024 | Official leaderboard |
| 41 | | 30.8% | Non-thinking | Aug 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | o3 | 92.6% | Thinking | Apr 2025 | Official leaderboard |
| 2 | GPT-5 mini | 91% | Thinking | Aug 2025 | Official leaderboard |
| 3 | GPT-5 nano | 89.7% | Thinking | Aug 2025 | Official leaderboard |
| 4 | | 89.1% | Thinking | Jun 2025 | Official leaderboard |
| 5 | gpt-oss-120b | 88.7% | Thinking | Aug 2025 | Official leaderboard |
| 6 | Claude 3.7 Sonnet | 88.1% | Non-thinking | Feb 2025 | Official leaderboard |
| 7 | | 88% | Thinking | Sep 2025 | Official leaderboard |
| 8 | GPT-4.1 | 86.3% | Non-thinking | Apr 2025 | Official leaderboard |
| 9 | gpt-oss-20b | 85.9% | Thinking | Aug 2025 | Official leaderboard |
| 10 | | 85% | Non-thinking | Feb 2025 | Official leaderboard |
| 11 | | 84.1% | Non-thinking | Feb 2025 | Official leaderboard |
| 12 | GPT-4.1 mini | 83.4% | Non-thinking | Apr 2025 | Official leaderboard |
| 13 | | 83.3% | Non-thinking | Apr 2025 | Official leaderboard |
| 14 | | 83% | Non-thinking | Mar 2025 | Official leaderboard |
| 15 | GPT-4o Mar 2025 | 83% | Non-thinking | Mar 2025 | Official leaderboard |
| 16 | Claude 3.5 Haiku | 82.5% | Non-thinking | Oct 2024 | Official leaderboard |
| 17 | | 81.5% | Non-thinking | Mar 2025 | Official leaderboard |
| 18 | | 81.1% | Non-thinking | Apr 2025 | Official leaderboard |
| 19 | | 80.3% | Non-thinking | Apr 2025 | Official leaderboard |
| 20 | GPT-4o mini | 80.2% | Non-thinking | Jul 2024 | Official leaderboard |
| 21 | | 79% | Non-thinking | Apr 2025 | Official leaderboard |
| 22 | GPT-4.1 nano | 75.8% | Non-thinking | Apr 2025 | Official leaderboard |
| 23 | | 75.8% | Non-thinking | Apr 2025 | Official leaderboard |
| 24 | | 75.3% | Non-thinking | Mar 2025 | Official leaderboard |
| 25 | | 75% | Non-thinking | Jun 2025 | Official leaderboard |
| 26 | | 74.7% | Non-thinking | Apr 2025 | Official leaderboard |
| 27 | | 74.5% | Non-thinking | Mar 2025 | Official leaderboard |
| 28 | | 71.3% | Non-thinking | Dec 2024 | Official leaderboard |
| 29 | | 70.1% | Non-thinking | Aug 2024 | Official leaderboard |
| 30 | | 69.9% | Non-thinking | Oct 2024 | Official leaderboard |
| 31 | deepseek-r1-0528-qwen3-8b | 68% | Thinking | May 2025 | Official leaderboard |
| 32 | | 57.9% | Non-thinking | Mar 2024 | Official leaderboard |
| 33 | | 54.5% | Non-thinking | Mar 2025 | Official leaderboard |
| 34 | | 53.3% | Non-thinking | Sep 2024 | Official leaderboard |
| 35 | | 36.6% | Non-thinking | Sep 2024 | Official leaderboard |
| 36 | | 30.8% | Non-thinking | Aug 2025 | Official leaderboard |
Rankings by part
We read 16 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as strict combination, strict detectable content and strict detectable format. o3 leads 5 of them, and 6 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
strict combination
36 models we track have a score here.
- 1 o3 85.9%
- 2
Gemini 2.5 Pro Jun 2025 ↑2 84.6%
- 3 Claude 3.7 Sonnet ↑3 83.1%
- 4 gpt-oss-120b ↑1 81.5%
- 5 GPT-4o Mar 2025 ↑10 80%
- 6
Gemini 2.5 Flash Sep 2025 ↑1 80%
- 7 GPT-4.1 ↑1 78.5%
- 8 GPT-4o mini ↑12 73.8%
- 9 Claude 3.5 Haiku ↑7 73.8%
- 10
Gemini 2.0 Flash-Lite ↑1 72.3%
strict detectable content
36 models we track have a score here.
- 1 GPT-5 mini ↑1 100%
- 2 Claude 3.7 Sonnet ↑4 97.8%
- 3 GPT-4o Mar 2025 ↑12 97.8%
- 4
Gemini 2.5 Pro Jun 2025 97.8%
- 5
Gemini 2.5 Flash Sep 2025 ↑2 97.7%
- 6 Claude 3.5 Haiku ↑10 95.7%
- 7 o3 ↓6 95.7%
- 8 gpt-oss-120b ↓3 95.7%
- 9 GPT-5 nano ↓6 95.7%
- 10 gpt-oss-20b ↓1 94.4%
strict detectable format
36 models we track have a score here.
- 1 gpt-oss-120b ↑4 95.9%
- 2 GPT-5 mini 92.5%
- 3
Qwen3-30B-A3B ↑10 91.4%
- 4 gpt-oss-20b ↑5 91.1%
- 5
Gemini 2.0 Flash Feb 2025 ↑5 90.5%
- 6
Qwen3-4B ↑17 90.5%
- 7 o3 ↓6 89.8%
- 8
Qwen3-14B ↑10 89.8%
- 9
Qwen3-32B ↑10 89.5%
- 10
Gemma 3 12B ↑7 89.1%
strict keywords
36 models we track have a score here.
- 1 o3 96.5%
- 2 GPT-5 mini 92.4%
- 3 GPT-5 nano 92.4%
- 4 Claude 3.7 Sonnet ↑2 89.7%
- 5 gpt-oss-20b ↑4 88%
- 6
Gemini 2.5 Pro Jun 2025 ↓2 87.6%
- 7 gpt-oss-120b ↓2 86.8%
- 8
Gemini 2.5 Flash Sep 2025 ↓1 86.2%
- 9 GPT-4.1 ↓1 84.8%
- 10 Claude 3.5 Haiku ↑6 84.1%
strict language
36 models we track have a score here.
- 1
Aya Expanse 32B ↑29 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑8 100%
- 3 Claude 3.7 Sonnet ↑3 100%
- 4
Gemma 3 27B ↑10 100%
- 5
Gemma 3 4B ↑22 100%
- 6 GPT-4o Mar 2025 ↑9 100%
- 7 GPT-4.1 ↑1 100%
- 8 o3 ↓7 100%
- 9
Gemma 3n E4B ↑16 100%
- 10 gpt-oss-120b ↓5 100%
strict length constraints
36 models we track have a score here.
- 1 GPT-5 mini ↑1 90.2%
- 2 GPT-5 nano ↑1 89.4%
- 3 o3 ↓2 88.5%
- 4
Gemini 2.5 Pro Jun 2025 82.1%
- 5
Gemini 2.5 Flash Sep 2025 ↑2 81.7%
- 6 gpt-oss-120b ↓1 79.7%
- 7
Gemma 3 27B ↑7 78.9%
- 8 GPT-4.1 78%
- 9
Gemini 2.0 Flash Feb 2025 ↑1 76.4%
- 10
Gemini 2.0 Flash-Lite ↑1 76.4%
strict punctuation
36 models we track have a score here.
- 1 Claude 3.5 Haiku ↑15 100%
- 2
Gemini 2.5 Pro Jun 2025 ↑2 100%
- 3
Gemma 3 27B ↑11 98.4%
- 4
Gemini 2.0 Flash Feb 2025 ↑6 96.7%
- 5
Gemini 2.0 Flash-Lite ↑6 96.7%
- 6
Gemini 2.5 Flash Sep 2025 ↑1 96.7%
- 7
Qwen3-30B-A3B ↑6 96.6%
- 8 Claude 3.7 Sonnet ↓2 95.1%
- 9 GPT-5 mini ↓7 95.1%
- 10 GPT-5 nano ↓7 95.1%
strict startend
36 models we track have a score here.
- 1 o3 96.8%
- 2 GPT-5 mini 96.8%
- 3
Gemini 2.5 Flash Sep 2025 ↑4 96.8%
- 4
Gemini 2.0 Flash Feb 2025 ↑6 95.2%
- 5
Gemini 2.0 Flash-Lite ↑6 95.2%
- 6 Claude 3.7 Sonnet 95.2%
- 7 GPT-4.1 ↑1 95.2%
- 8 GPT-4.1 mini ↑4 95.2%
- 9
Gemini 2.5 Pro Jun 2025 ↓5 95.2%
- 10 gpt-oss-20b ↓1 93.8%
loose combination
36 models we track have a score here.
- 1 o3 87.5%
- 2 Claude 3.7 Sonnet ↑4 84.6%
- 3
Gemini 2.5 Pro Jun 2025 ↑1 84.6%
- 4 gpt-oss-120b ↑1 84.6%
- 5
Gemini 2.5 Flash Sep 2025 ↑2 83.1%
- 6 GPT-4o Mar 2025 ↑9 81.5%
- 7 GPT-4.1 ↑1 81.5%
- 8
Qwen3-30B-A3B ↑5 79.4%
- 9 gpt-oss-20b 77.8%
- 10
Qwen3-14B ↑8 76.9%
loose detectable content
36 models we track have a score here.
- 1 GPT-5 mini ↑1 100%
- 2 Claude 3.7 Sonnet ↑4 97.8%
- 3 GPT-4o Mar 2025 ↑12 97.8%
- 4
Gemini 2.5 Pro Jun 2025 97.8%
- 5
Gemini 2.5 Flash Sep 2025 ↑2 97.7%
- 6 Claude 3.5 Haiku ↑10 95.7%
- 7 o3 ↓6 95.7%
- 8 gpt-oss-120b ↓3 95.7%
- 9 GPT-5 nano ↓6 95.7%
- 10 gpt-oss-20b ↓1 94.4%
loose detectable format
36 models we track have a score here.
- 1 gpt-oss-120b ↑4 96.6%
- 2
Qwen3-30B-A3B ↑11 92.8%
- 3 GPT-5 mini ↓1 92.5%
- 4 GPT-4o Mar 2025 ↑11 91.2%
- 5
Qwen3-14B ↑13 91.2%
- 6
Qwen3-4B ↑17 91.2%
- 7 gpt-oss-20b ↑2 91.1%
- 8
Qwen3-32B ↑11 90.9%
- 9
Gemini 2.0 Flash Feb 2025 ↑1 90.5%
- 10 GPT-4.1 ↓2 90.5%
loose keywords
36 models we track have a score here.
- 1 o3 97.9%
- 2 GPT-5 mini 95.2%
- 3 Claude 3.7 Sonnet ↑3 94.5%
- 4 GPT-5 nano ↓1 94.5%
- 5 gpt-oss-20b ↑4 92.4%
- 6 gpt-oss-120b ↓1 92.4%
- 7
Gemini 2.5 Pro Jun 2025 ↓3 91.7%
- 8
Gemini 2.5 Flash Sep 2025 ↓1 91.3%
- 9 GPT-4.1 ↓1 91%
- 10 GPT-4.1 mini ↑2 88.3%
loose language
36 models we track have a score here.
- 1
Aya Expanse 32B ↑29 100%
- 2
Gemini 2.0 Flash Feb 2025 ↑8 100%
- 3
Gemini 2.0 Flash-Lite ↑8 100%
- 4 Claude 3.7 Sonnet ↑2 100%
- 5
Gemma 3 27B ↑9 100%
- 6
Gemma 3 12B ↑11 100%
- 7
Gemma 3 4B ↑20 100%
- 8 GPT-4o Mar 2025 ↑7 100%
- 9 GPT-4.1 ↓1 100%
- 10 o3 ↓9 100%
loose length constraints
36 models we track have a score here.
- 1 GPT-5 nano ↑2 94.3%
- 2
Gemini 2.5 Pro Jun 2025 ↑2 92.7%
- 3 GPT-5 mini ↓1 91.1%
- 4 o3 ↓3 90.2%
- 5
Gemma 3 27B ↑9 87.8%
- 6
Gemini 2.5 Flash Sep 2025 ↑1 86.7%
- 7 gpt-oss-120b ↓2 86.2%
- 8 Claude 3.7 Sonnet ↓2 85.4%
- 9
Gemini 2.0 Flash Feb 2025 ↑1 83.7%
- 10 GPT-4.1 ↓2 83.7%
loose punctuation
36 models we track have a score here.
- 1 Claude 3.5 Haiku ↑15 100%
- 2 o3 ↓1 100%
- 3
Gemini 2.5 Pro Jun 2025 ↑1 100%
- 4 GPT-5 mini ↓2 100%
- 5 GPT-5 nano ↓2 100%
- 6
Gemini 2.5 Flash Sep 2025 ↑1 100%
- 7 Claude 3.7 Sonnet ↓1 98.4%
- 8
Gemma 3 27B ↑6 98.4%
- 9
Gemini 2.0 Flash Feb 2025 ↑1 96.7%
- 10
Gemini 2.0 Flash-Lite ↑1 96.7%
loose startend
36 models we track have a score here.
- 1 Claude 3.7 Sonnet ↑5 96.8%
- 2 o3 ↓1 96.8%
- 3
Qwen3-32B ↑16 96.8%
- 4 GPT-5 mini ↓2 96.8%
- 5
Gemini 2.5 Flash Sep 2025 ↑2 96.8%
- 6
Gemini 2.0 Flash Feb 2025 ↑4 95.2%
- 7
Gemini 2.0 Flash-Lite ↑4 95.2%
- 8 GPT-4.1 95.2%
- 9 GPT-4.1 mini ↑3 95.2%
- 10
Qwen3-14B ↑8 95.2%
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Persian IFEval leaderboard
Claude 3.7 Sonnet on the official leaderboard passed 85% in February 2025
Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Persian IFEval 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
All 8 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the official leaderboard (41 entries).
- Pass Models scored 36 models we track, plus 5 other entries on its leaderboard.
- Note Newest score The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash. That’s expected for a saturated test.
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is o3 (#84).
- Note Headroom Saturated. The best is 92.6%, past 85% since February 2025.
Leaderboard import 3 pass
- Pass Importer We read the board on mcinext-mizan-llm-leaderboard.hf.space: 40 rows for models we track and 5 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.