Benchmark Almanac

Persian IFEval

o3 92.6% leads since May 2025
Not in model cards

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
In and out
Text → Text
Interaction
Single turn
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
41
Entries on its own leaderboard, 36 models we track

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:

  1. 1 o3 Thinking 92.6%
  2. 2 GPT-5 mini Thinking 91%
  3. 3 GPT-5 nano Thinking 89.7%
  4. 4 Gemini 2.5 Pro Jun 2025, Thinking 89.1%
  5. 5 gpt-oss-120b Thinking 88.7%
  6. 6 Claude 3.7 Sonnet Non-thinking 88.1%
  7. 7 Gemini 2.5 Flash Sep 2025, Thinking 88%
  8. 8 GPT-4.1 Non-thinking 86.3%
  9. 9 gpt-oss-20b Thinking 85.9%
  10. 10 Gemini 2.0 Flash Feb 2025, Non-thinking 85%
  11. 11 c4ai-command-a-03-2025 Non-thinking 84.4%
  12. 12 Gemini 2.0 Flash-Lite Non-thinking 84.1%
  13. 13 deepseek-reasoner Thinking 83.7%
  14. 14 GPT-4.1 mini Non-thinking 83.4%
  15. 15 Qwen3-30B-A3B Non-thinking 83.3%
  16. 16 deepseek-chat Non-thinking 83.1%
  17. 17 Gemma 3 27B Non-thinking 83%
  18. 18 GPT-4o Mar 2025, Non-thinking 83%
  19. 19 Claude 3.5 Haiku Non-thinking 82.5%
  20. 20 Gemma 3 12B Non-thinking 81.5%
  21. 21 Qwen3-14B Non-thinking 81.1%
  22. 22 Qwen3-32B Non-thinking 80.3%
  23. 23 GPT-4o mini Non-thinking 80.2%
  24. 24 Llama 4 Scout Non-thinking 79%
  25. 25 GPT-4.1 nano Non-thinking 75.8%
  26. 26 Qwen3-4B Non-thinking 75.8%
  27. 27 Mistral Small 3.1 Non-thinking 75.3%
  28. 28 Gemma 3n E4B Non-thinking 75%
  29. 29 Qwen3-8B Non-thinking 74.7%
  30. 30 Gemma 3 4B Non-thinking 74.5%
  31. 31 Llama 3.3 70B Instruct Non-thinking 71.3%
  32. 32 Command R+ Non-thinking 70.1%
  33. 33 Aya Expanse 32B Non-thinking 69.9%
  34. 34 deepseek-r1-0528-qwen3-8b Thinking 68%
  35. 35 yandexgpt-lite Non-thinking 60.4%
  36. 36 Command R (03-2024) Non-thinking 57.9%
  37. 37 Gemma 3 1B Non-thinking 54.5%
  38. 38 Llama 3.2 3B Non-thinking 53.3%
  39. 39 Mistral-7B-Instruct-v0.3 Non-thinking 44%
  40. 40 Llama 3.2 1B Non-thinking 36.6%
  41. 41 Gemma 3 270M Non-thinking 30.8%

Source: Official Persian IFEval leaderboard

Source: Official Persian IFEval leaderboard

#EntryScoreSettingDateSource
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 Gemini 2.5 Pro Jun 2025 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 Gemini 2.5 Flash Sep 2025 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 Gemini 2.0 Flash Feb 2025 85% Non-thinking Feb 2025 Official leaderboard
11 c4ai-command-a-03-2025 open weights, 111B 84.4% Non-thinking Official leaderboard
12 Gemini 2.0 Flash-Lite 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 Qwen3-30B-A3B 83.3% Non-thinking Apr 2025 Official leaderboard
16 deepseek-chat open weights, 671B 83.1% Non-thinking Official leaderboard
17 Gemma 3 27B 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 Gemma 3 12B 81.5% Non-thinking Mar 2025 Official leaderboard
21 Qwen3-14B 81.1% Non-thinking Apr 2025 Official leaderboard
22 Qwen3-32B 80.3% Non-thinking Apr 2025 Official leaderboard
23 GPT-4o mini 80.2% Non-thinking Jul 2024 Official leaderboard
24 Llama 4 Scout 79% Non-thinking Apr 2025 Official leaderboard
25 GPT-4.1 nano 75.8% Non-thinking Apr 2025 Official leaderboard
26 Qwen3-4B 75.8% Non-thinking Apr 2025 Official leaderboard
27 Mistral Small 3.1 75.3% Non-thinking Mar 2025 Official leaderboard
28 Gemma 3n E4B 75% Non-thinking Jun 2025 Official leaderboard
29 Qwen3-8B 74.7% Non-thinking Apr 2025 Official leaderboard
30 Gemma 3 4B 74.5% Non-thinking Mar 2025 Official leaderboard
31 Llama 3.3 70B Instruct 71.3% Non-thinking Dec 2024 Official leaderboard
32 Command R+ 70.1% Non-thinking Aug 2024 Official leaderboard
33 Aya Expanse 32B 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 Command R (03-2024) 57.9% Non-thinking Mar 2024 Official leaderboard
37 Gemma 3 1B 54.5% Non-thinking Mar 2025 Official leaderboard
38 Llama 3.2 3B 53.3% Non-thinking Sep 2024 Official leaderboard
39 Mistral-7B-Instruct-v0.3 open weights, 7.3B 44% Non-thinking Official leaderboard
40 Llama 3.2 1B 36.6% Non-thinking Sep 2024 Official leaderboard
41 Gemma 3 270M 30.8% Non-thinking Aug 2025 Official leaderboard
#ModelScoreSettingDateSource
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 Gemini 2.5 Pro Jun 2025 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 Gemini 2.5 Flash Sep 2025 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 Gemini 2.0 Flash Feb 2025 85% Non-thinking Feb 2025 Official leaderboard
11 Gemini 2.0 Flash-Lite 84.1% Non-thinking Feb 2025 Official leaderboard
12 GPT-4.1 mini 83.4% Non-thinking Apr 2025 Official leaderboard
13 Qwen3-30B-A3B 83.3% Non-thinking Apr 2025 Official leaderboard
14 Gemma 3 27B 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 Gemma 3 12B 81.5% Non-thinking Mar 2025 Official leaderboard
18 Qwen3-14B 81.1% Non-thinking Apr 2025 Official leaderboard
19 Qwen3-32B 80.3% Non-thinking Apr 2025 Official leaderboard
20 GPT-4o mini 80.2% Non-thinking Jul 2024 Official leaderboard
21 Llama 4 Scout 79% Non-thinking Apr 2025 Official leaderboard
22 GPT-4.1 nano 75.8% Non-thinking Apr 2025 Official leaderboard
23 Qwen3-4B 75.8% Non-thinking Apr 2025 Official leaderboard
24 Mistral Small 3.1 75.3% Non-thinking Mar 2025 Official leaderboard
25 Gemma 3n E4B 75% Non-thinking Jun 2025 Official leaderboard
26 Qwen3-8B 74.7% Non-thinking Apr 2025 Official leaderboard
27 Gemma 3 4B 74.5% Non-thinking Mar 2025 Official leaderboard
28 Llama 3.3 70B Instruct 71.3% Non-thinking Dec 2024 Official leaderboard
29 Command R+ 70.1% Non-thinking Aug 2024 Official leaderboard
30 Aya Expanse 32B 69.9% Non-thinking Oct 2024 Official leaderboard
31 deepseek-r1-0528-qwen3-8b 68% Thinking May 2025 Official leaderboard
32 Command R (03-2024) 57.9% Non-thinking Mar 2024 Official leaderboard
33 Gemma 3 1B 54.5% Non-thinking Mar 2025 Official leaderboard
34 Llama 3.2 3B 53.3% Non-thinking Sep 2024 Official leaderboard
35 Llama 3.2 1B 36.6% Non-thinking Sep 2024 Official leaderboard
36 Gemma 3 270M 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. 1 o3 85.9%
  2. 2 Gemini 2.5 Pro Jun 2025 ↑2 84.6%
  3. 3 Claude 3.7 Sonnet ↑3 83.1%
  4. 4 gpt-oss-120b ↑1 81.5%
  5. 5 GPT-4o Mar 2025 ↑10 80%
  6. 6 Gemini 2.5 Flash Sep 2025 ↑1 80%
  7. 7 GPT-4.1 ↑1 78.5%
  8. 8 GPT-4o mini ↑12 73.8%
  9. 9 Claude 3.5 Haiku ↑7 73.8%
  10. 10 Gemini 2.0 Flash-Lite ↑1 72.3%

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI
Own leaderboard Passed 85% in May 2025, by o3
o3 92.6%

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.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) o3 Claude 3.7 Sonnet Claude 3.5 Haiku Command R (03-2024) 0% 20% 40% 60% 80% 100% 2024 2025 Near the ceiling (85%) o3 Gemini 2.0 Flash

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

All 545 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.

How we check