Persian MT-Bench
Models follow Persian conversations across two to four turns. One asks for film directors, then actors, then the oldest and newest films.
MIZAN ranks 45 model configurations on 100 dialogues. Ten dialogues cover Iranian cultural knowledge and ten test retrieval during conversation.
- Made by
- MCINext
- Released
- May 2025
- Kind
- Chat and instructions in many languages
- Level
- Everyday
- Scored
- Graded by a model
- Answers
- Open-ended
- Tasks
- All public
- Measures
- Capability
- Language
- Persian
- Interaction
- Multi-turn
- Items from
- Another benchmark
- Refresh
- One fixed set
- Run by
- Anyone
- Based on
- MT-Bench
- Strongest model tested
- o3 #84 on the capability index
- Last model added
- September 2025
- Scores refreshed
- November 20th, 2026
o3 leads with 9 among models released by September 2025 o3 leads with 9 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. Every run is in the next chart. 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
- 9 o3
- 9
Gemini 2.5 Pro
- 9
Gemini 2.5 Flash
- 9 GPT-4.1
- 9 Claude 3.7 Sonnet
- 9 GPT-5 mini
- 9 deepseek-reasoner
- 9 deepseek-chat
- 8 gpt-oss-120b
- 8 GPT-4.1 mini
- 8
Gemini 2.0 Flash
- 8 GPT-4o
- 8 GPT-5 nano
- 8
Gemini 2.0 Flash-Lite
- 8
Cohere Command A
- 8 gpt-oss-20b
- 8
Gemma 3 27B
- 8 GPT-4o mini
- 8
Qwen3-32B
- 8 Claude 3.5 Haiku
- 8
Gemma 3 12B
- 7
Qwen3-30B-A3B
- 7
Llama 4 Scout
- 7 GPT-4.1 nano
- 7
Mistral Small 3.1
- 7
Qwen3-14B
- 7
Llama 3.3 70B Instruct
- 7
Aya Expanse 32B
- 7
Command R+
- 7
Gemma 3n E4B
- 7 deepseek-r1-0528-qwen3-8b
- 7
Gemma 3 4B
- 7
Qwen3-8B
- 6
Command R (03-2024)
- 6
Qwen3-4B
- 5 yandexgpt-lite
- 4
Gemma 3 1B
- 4 Llama-3.2-3B-Instruct
- 3 Mistral-7B-Instruct-v0.3
- 3 Llama-3.2-1B-Instruct
- 2
Gemma 3 270M
- 1 o3 Thinking 9
- 2
Gemini 2.5 Pro Jun 2025, Thinking 9
- 3
Gemini 2.5 Flash Sep 2025, Thinking 9
- 4 GPT-4.1 Non-thinking 9
- 5 Claude 3.7 Sonnet Non-thinking 9
- 6 GPT-5 mini Thinking 9
- 7 deepseek-reasoner Thinking 9
- 8 deepseek-chat Non-thinking 9
- 9 gpt-oss-120b Thinking 8
- 10 GPT-4.1 mini Non-thinking 8
- 11
Gemini 2.0 Flash Feb 2025, Non-thinking 8
- 12 GPT-4o Mar 2025, Non-thinking 8
- 13 GPT-5 nano Thinking 8
- 14
Gemini 2.0 Flash-Lite Non-thinking 8
- 15
Cohere Command A Non-thinking 8
- 16 gpt-oss-20b Thinking 8
- 17
Gemma 3 27B Non-thinking 8
- 18 GPT-4o mini Non-thinking 8
- 19
Qwen3-32B Non-thinking 8
- 20 Claude 3.5 Haiku Non-thinking 8
- 21
Gemma 3 12B Non-thinking 8
- 22
Qwen3-30B-A3B Non-thinking 7
- 23
Llama 4 Scout Non-thinking 7
- 24 GPT-4.1 nano Non-thinking 7
- 25
Mistral Small 3.1 Non-thinking 7
- 26
Qwen3-14B Non-thinking 7
- 27
Llama 3.3 70B Instruct Non-thinking 7
- 28
Aya Expanse 32B Non-thinking 7
- 29
Command R+ Non-thinking 7
- 30
Gemma 3n E4B Non-thinking 7
- 31 deepseek-r1-0528-qwen3-8b Thinking 7
- 32
Gemma 3 4B Non-thinking 7
- 33
Qwen3-8B Non-thinking 7
- 34
Command R (03-2024) Non-thinking 6
- 35
Qwen3-4B Non-thinking 6
- 36 yandexgpt-lite Non-thinking 5
- 37
Gemma 3 1B Non-thinking 4
- 38 Llama-3.2-3B-Instruct Non-thinking 4
- 39 Mistral-7B-Instruct-v0.3 Non-thinking 3
- 40 Llama-3.2-1B-Instruct Non-thinking 3
- 41
Gemma 3 270M Non-thinking 2
Source: Official Persian MT-Bench leaderboard
- 9 o3
- 9
Gemini 2.5 Pro
- 9
Gemini 2.5 Flash
- 9 GPT-4.1
- 9 Claude 3.7 Sonnet
- 9 GPT-5 mini
- 8 gpt-oss-120b
- 8 GPT-4.1 mini
- 8
Gemini 2.0 Flash
- 8 GPT-4o
- 8 GPT-5 nano
- 8
Gemini 2.0 Flash-Lite
- 8
Cohere Command A
- 8 gpt-oss-20b
- 8
Gemma 3 27B
- 8 GPT-4o mini
- 8
Qwen3-32B
- 8 Claude 3.5 Haiku
- 8
Gemma 3 12B
- 7
Qwen3-30B-A3B
- 7
Llama 4 Scout
- 7 GPT-4.1 nano
- 7
Mistral Small 3.1
- 7
Qwen3-14B
- 7
Llama 3.3 70B Instruct
- 7
Aya Expanse 32B
- 7
Command R+
- 7
Gemma 3n E4B
- 7 deepseek-r1-0528-qwen3-8b
- 7
Gemma 3 4B
- 7
Qwen3-8B
- 6
Command R (03-2024)
- 6
Qwen3-4B
- 4
Gemma 3 1B
- 2
Gemma 3 270M
- 1 o3 Thinking 9
- 2
Gemini 2.5 Pro Jun 2025, Thinking 9
- 3
Gemini 2.5 Flash Sep 2025, Thinking 9
- 4 GPT-4.1 Non-thinking 9
- 5 Claude 3.7 Sonnet Non-thinking 9
- 6 GPT-5 mini Thinking 9
- 7 gpt-oss-120b Thinking 8
- 8 GPT-4.1 mini Non-thinking 8
- 9
Gemini 2.0 Flash Feb 2025, Non-thinking 8
- 10 GPT-4o Mar 2025, Non-thinking 8
- 11 GPT-5 nano Thinking 8
- 12
Gemini 2.0 Flash-Lite Non-thinking 8
- 13
Cohere Command A Non-thinking 8
- 14 gpt-oss-20b Thinking 8
- 15
Gemma 3 27B Non-thinking 8
- 16 GPT-4o mini Non-thinking 8
- 17
Qwen3-32B Non-thinking 8
- 18 Claude 3.5 Haiku Non-thinking 8
- 19
Gemma 3 12B Non-thinking 8
- 20
Qwen3-30B-A3B Non-thinking 7
- 21
Llama 4 Scout Non-thinking 7
- 22 GPT-4.1 nano Non-thinking 7
- 23
Mistral Small 3.1 Non-thinking 7
- 24
Qwen3-14B Non-thinking 7
- 25
Llama 3.3 70B Instruct Non-thinking 7
- 26
Aya Expanse 32B Non-thinking 7
- 27
Command R+ Non-thinking 7
- 28
Gemma 3n E4B Non-thinking 7
- 29 deepseek-r1-0528-qwen3-8b Thinking 7
- 30
Gemma 3 4B Non-thinking 7
- 31
Qwen3-8B Non-thinking 7
- 32
Command R (03-2024) Non-thinking 6
- 33
Qwen3-4B Non-thinking 6
- 34
Gemma 3 1B Non-thinking 4
- 35
Gemma 3 270M Non-thinking 2
Source: Official Persian MT-Bench leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | o3 | 9 | Thinking | Apr 2025 | Official leaderboard |
| 2 | | 9 | Thinking | Jun 2025 | Official leaderboard |
| 3 | | 9 | Thinking | Sep 2025 | Official leaderboard |
| 4 | GPT-4.1 | 9 | Non-thinking | Apr 2025 | Official leaderboard |
| 5 | Claude 3.7 Sonnet | 9 | Non-thinking | Feb 2025 | Official leaderboard |
| 6 | GPT-5 mini | 9 | Thinking | Aug 2025 | Official leaderboard |
| 7 | deepseek-reasoner open weights, 671B | 9 | Thinking | Official leaderboard | |
| 8 | deepseek-chat open weights, 671B | 9 | Non-thinking | Official leaderboard | |
| 9 | gpt-oss-120b | 8 | Thinking | Aug 2025 | Official leaderboard |
| 10 | GPT-4.1 mini | 8 | Non-thinking | Apr 2025 | Official leaderboard |
| 11 | | 8 | Non-thinking | Feb 2025 | Official leaderboard |
| 12 | GPT-4o Mar 2025 | 8 | Non-thinking | Mar 2025 | Official leaderboard |
| 13 | GPT-5 nano | 8 | Thinking | Aug 2025 | Official leaderboard |
| 14 | | 8 | Non-thinking | Feb 2025 | Official leaderboard |
| 15 | | 8 | Non-thinking | Mar 2025 | Official leaderboard |
| 16 | gpt-oss-20b | 8 | Thinking | Aug 2025 | Official leaderboard |
| 17 | | 8 | Non-thinking | Mar 2025 | Official leaderboard |
| 18 | GPT-4o mini | 8 | Non-thinking | Jul 2024 | Official leaderboard |
| 19 | | 8 | Non-thinking | Apr 2025 | Official leaderboard |
| 20 | Claude 3.5 Haiku | 8 | Non-thinking | Oct 2024 | Official leaderboard |
| 21 | | 8 | Non-thinking | Mar 2025 | Official leaderboard |
| 22 | | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 23 | | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 24 | GPT-4.1 nano | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 25 | | 7 | Non-thinking | Mar 2025 | Official leaderboard |
| 26 | | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 27 | | 7 | Non-thinking | Dec 2024 | Official leaderboard |
| 28 | | 7 | Non-thinking | Oct 2024 | Official leaderboard |
| 29 | | 7 | Non-thinking | Aug 2024 | Official leaderboard |
| 30 | | 7 | Non-thinking | Jun 2025 | Official leaderboard |
| 31 | deepseek-r1-0528-qwen3-8b | 7 | Thinking | May 2025 | Official leaderboard |
| 32 | | 7 | Non-thinking | Mar 2025 | Official leaderboard |
| 33 | | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 34 | | 6 | Non-thinking | Mar 2024 | Official leaderboard |
| 35 | | 6 | Non-thinking | Apr 2025 | Official leaderboard |
| 36 | yandexgpt-lite open weights, 8B | 5 | Non-thinking | Official leaderboard | |
| 37 | | 4 | Non-thinking | Mar 2025 | Official leaderboard |
| 38 | Llama-3.2-3B-Instruct open weights, 3.2B | 4 | Non-thinking | Official leaderboard | |
| 39 | Mistral-7B-Instruct-v0.3 open weights, 7.3B | 3 | Non-thinking | Official leaderboard | |
| 40 | Llama-3.2-1B-Instruct open weights, 1.2B | 3 | Non-thinking | Official leaderboard | |
| 41 | | 2 | Non-thinking | Aug 2025 | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | o3 | 9 | Thinking | Apr 2025 | Official leaderboard |
| 2 | | 9 | Thinking | Jun 2025 | Official leaderboard |
| 3 | | 9 | Thinking | Sep 2025 | Official leaderboard |
| 4 | GPT-4.1 | 9 | Non-thinking | Apr 2025 | Official leaderboard |
| 5 | Claude 3.7 Sonnet | 9 | Non-thinking | Feb 2025 | Official leaderboard |
| 6 | GPT-5 mini | 9 | Thinking | Aug 2025 | Official leaderboard |
| 7 | gpt-oss-120b | 8 | Thinking | Aug 2025 | Official leaderboard |
| 8 | GPT-4.1 mini | 8 | Non-thinking | Apr 2025 | Official leaderboard |
| 9 | | 8 | Non-thinking | Feb 2025 | Official leaderboard |
| 10 | GPT-4o Mar 2025 | 8 | Non-thinking | Mar 2025 | Official leaderboard |
| 11 | GPT-5 nano | 8 | Thinking | Aug 2025 | Official leaderboard |
| 12 | | 8 | Non-thinking | Feb 2025 | Official leaderboard |
| 13 | | 8 | Non-thinking | Mar 2025 | Official leaderboard |
| 14 | gpt-oss-20b | 8 | Thinking | Aug 2025 | Official leaderboard |
| 15 | | 8 | Non-thinking | Mar 2025 | Official leaderboard |
| 16 | GPT-4o mini | 8 | Non-thinking | Jul 2024 | Official leaderboard |
| 17 | | 8 | Non-thinking | Apr 2025 | Official leaderboard |
| 18 | Claude 3.5 Haiku | 8 | Non-thinking | Oct 2024 | Official leaderboard |
| 19 | | 8 | Non-thinking | Mar 2025 | Official leaderboard |
| 20 | | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 21 | | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 22 | GPT-4.1 nano | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 23 | | 7 | Non-thinking | Mar 2025 | Official leaderboard |
| 24 | | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 25 | | 7 | Non-thinking | Dec 2024 | Official leaderboard |
| 26 | | 7 | Non-thinking | Oct 2024 | Official leaderboard |
| 27 | | 7 | Non-thinking | Aug 2024 | Official leaderboard |
| 28 | | 7 | Non-thinking | Jun 2025 | Official leaderboard |
| 29 | deepseek-r1-0528-qwen3-8b | 7 | Thinking | May 2025 | Official leaderboard |
| 30 | | 7 | Non-thinking | Mar 2025 | Official leaderboard |
| 31 | | 7 | Non-thinking | Apr 2025 | Official leaderboard |
| 32 | | 6 | Non-thinking | Mar 2024 | Official leaderboard |
| 33 | | 6 | Non-thinking | Apr 2025 | Official leaderboard |
| 34 | | 4 | Non-thinking | Mar 2025 | Official leaderboard |
| 35 | | 2 | Non-thinking | Aug 2025 | Official leaderboard |
Rankings by part
We read 21 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as score w mean, writing score w mean and writing score mean. o3 leads 9 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.
score w mean
35 models we track have a score here.
- 1 o3 9
- 2
Gemini 2.5 Pro Jun 2025 9
- 3
Gemini 2.5 Flash Sep 2025 9
- 4 Claude 3.7 Sonnet ↑1 9
- 5 GPT-4.1 ↓1 9
- 6 GPT-5 mini 9
- 7
Gemini 2.0 Flash Feb 2025 ↑2 8
- 8 gpt-oss-120b ↓1 8
- 9 GPT-4o Mar 2025 ↑1 8
- 10 GPT-4.1 mini ↓2 8
writing score w mean
35 models we track have a score here.
- 1 o3 9
- 2
Gemini 2.5 Pro Jun 2025 9
- 3 Claude 3.7 Sonnet ↑2 9
- 4
Gemini 2.5 Flash Sep 2025 ↓1 9
- 5
Gemini 2.0 Flash Feb 2025 ↑4 8
- 6
Gemini 2.0 Flash-Lite ↑6 8
- 7 GPT-4.1 ↓3 8
- 8
Gemma 3 27B ↑7 8
- 9 gpt-oss-120b ↓2 8
- 10 GPT-5 mini ↓4 8
writing score mean
35 models we track have a score here.
- 1 o3 9
- 2
Gemini 2.5 Pro Jun 2025 9
- 3 Claude 3.7 Sonnet ↑2 9
- 4
Gemini 2.5 Flash Sep 2025 ↓1 9
- 5
Gemini 2.0 Flash Feb 2025 ↑4 8
- 6 GPT-4.1 ↓2 8
- 7
Gemini 2.0 Flash-Lite ↑5 8
- 8
Gemma 3 27B ↑7 8
- 9
Cohere Command A ↑4 8
- 10 GPT-5 mini ↓4 8
roleplay score w mean
35 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑1 9
- 2
Gemini 2.0 Flash Feb 2025 ↑7 9
- 3 GPT-4.1 ↑1 9
- 4 o3 ↓3 9
- 5 GPT-5 mini ↑1 8
- 6
Cohere Command A ↑7 8
- 7
Gemini 2.5 Flash Sep 2025 ↓4 8
- 8 Claude 3.7 Sonnet ↓3 8
- 9 GPT-4.1 mini ↓1 8
- 10
Gemini 2.0 Flash-Lite ↑2 8
roleplay score mean
35 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑1 9
- 2
Gemini 2.0 Flash Feb 2025 ↑7 9
- 3 GPT-4.1 ↑1 9
- 4 o3 ↓3 9
- 5 GPT-5 mini ↑1 8
- 6
Cohere Command A ↑7 8
- 7
Gemini 2.5 Flash Sep 2025 ↓4 8
- 8 Claude 3.7 Sonnet ↓3 8
- 9 GPT-4.1 mini ↓1 8
- 10
Gemini 2.0 Flash-Lite ↑2 8
reasoning score w mean
35 models we track have a score here.
- 1 o3 10
- 2 Claude 3.7 Sonnet ↑3 9
- 3 GPT-4.1 ↑1 9
- 4
Gemini 2.5 Pro Jun 2025 ↓2 9
- 5 gpt-oss-120b ↑2 9
- 6 GPT-5 mini 9
- 7 GPT-4o Mar 2025 ↑3 8
- 8
Gemini 2.5 Flash Sep 2025 ↓5 8
- 9 GPT-5 nano ↑2 8
- 10
Gemini 2.0 Flash Feb 2025 ↓1 8
reasoning score mean
35 models we track have a score here.
- 1 o3 10
- 2 GPT-5 mini ↑4 9
- 3 GPT-4.1 ↑1 9
- 4
Gemini 2.5 Pro Jun 2025 ↓2 9
- 5 gpt-oss-120b ↑2 9
- 6 Claude 3.7 Sonnet ↓1 9
- 7 GPT-4o Mar 2025 ↑3 8
- 8
Gemini 2.5 Flash Sep 2025 ↓5 8
- 9
Gemini 2.0 Flash Feb 2025 8
- 10 GPT-5 nano ↑1 8
math score w mean
35 models we track have a score here.
- 1 Claude 3.7 Sonnet ↑4 10
- 2 GPT-4.1 ↑2 10
- 3
Gemini 2.5 Pro Jun 2025 ↓1 10
- 4 gpt-oss-20b ↑10 10
- 5 o3 ↓4 10
- 6 deepseek-r1-0528-qwen3-8b ↑23 10
- 7 GPT-5 mini ↓1 10
- 8 GPT-5 nano ↑3 10
- 9
Gemini 2.5 Flash Sep 2025 ↓6 10
- 10
Qwen3-30B-A3B ↑10 10
math score mean
35 models we track have a score here.
- 1 Claude 3.7 Sonnet ↑4 10
- 2 GPT-4.1 ↑2 10
- 3
Gemini 2.5 Pro Jun 2025 ↓1 10
- 4 gpt-oss-20b ↑10 10
- 5 o3 ↓4 10
- 6 deepseek-r1-0528-qwen3-8b ↑23 10
- 7 GPT-5 mini ↓1 10
- 8 GPT-5 nano ↑3 10
- 9
Gemini 2.5 Flash Sep 2025 ↓6 10
- 10
Qwen3-30B-A3B ↑10 10
coding score w mean
35 models we track have a score here.
- 1 gpt-oss-120b ↑6 9
- 2
Gemini 2.5 Pro Jun 2025 9
- 3 gpt-oss-20b ↑11 9
- 4 GPT-5 mini ↑2 9
- 5 o3 ↓4 9
- 6 GPT-5 nano ↑5 9
- 7 GPT-4.1 mini ↑1 9
- 8
Gemini 2.5 Flash Sep 2025 ↓5 9
- 9
Qwen3-30B-A3B ↑11 9
- 10
Gemini 2.0 Flash-Lite ↑2 8
coding score mean
35 models we track have a score here.
- 1 gpt-oss-120b ↑6 9
- 2
Gemini 2.5 Pro Jun 2025 9
- 3 gpt-oss-20b ↑11 9
- 4 GPT-5 mini ↑2 9
- 5 o3 ↓4 9
- 6 GPT-5 nano ↑5 9
- 7 GPT-4.1 mini ↑1 9
- 8
Gemini 2.5 Flash Sep 2025 ↓5 9
- 9
Qwen3-30B-A3B ↑11 9
- 10
Gemini 2.0 Flash-Lite ↑2 8
extraction score w mean
34 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑1 9
- 2 GPT-5 mini ↑4 9
- 3 o3 ↓2 9
- 4
Gemini 2.5 Flash Sep 2025 ↓1 9
- 5 Claude 3.7 Sonnet 9
- 6 GPT-4.1 mini ↑2 9
- 7 gpt-oss-120b 9
- 8 GPT-4.1 ↓4 8
- 9
Gemma 3 27B ↑6 8
- 10 gpt-oss-20b ↑4 8
extraction score mean
34 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑1 9
- 2 GPT-5 mini ↑4 9
- 3 o3 ↓2 9
- 4
Gemini 2.5 Flash Sep 2025 ↓1 9
- 5 Claude 3.7 Sonnet 9
- 6 GPT-4.1 mini ↑2 9
- 7 gpt-oss-120b 9
- 8 GPT-4.1 ↓4 8
- 9
Gemma 3 27B ↑6 8
- 10 gpt-oss-20b ↑4 8
stem score w mean
35 models we track have a score here.
- 1 o3 9
- 2 gpt-oss-120b ↑5 9
- 3
Qwen3-32B ↑14 9
- 4 GPT-5 mini ↑2 9
- 5 GPT-4.1 mini ↑3 9
- 6 gpt-oss-20b ↑8 9
- 7
Gemini 2.5 Pro Jun 2025 ↓5 9
- 8 GPT-4.1 ↓4 9
- 9
Gemini 2.5 Flash Sep 2025 ↓6 9
- 10
Gemma 3 27B ↑5 9
stem score mean
35 models we track have a score here.
- 1 o3 9
- 2 gpt-oss-120b ↑5 9
- 3
Qwen3-32B ↑14 9
- 4 GPT-5 mini ↑2 9
- 5 GPT-4.1 mini ↑3 9
- 6 gpt-oss-20b ↑8 9
- 7
Gemini 2.5 Pro Jun 2025 ↓5 9
- 8 GPT-4.1 ↓4 9
- 9
Gemini 2.5 Flash Sep 2025 ↓6 9
- 10
Gemma 3 27B ↑5 9
humanities score w mean
35 models we track have a score here.
- 1 gpt-oss-120b ↑6 9
- 2 Claude 3.7 Sonnet ↑3 9
- 3 GPT-5 mini ↑3 9
- 4
Gemini 2.5 Pro Jun 2025 ↓2 9
- 5 GPT-4.1 ↓1 9
- 6 o3 ↓5 9
- 7 GPT-4o Mar 2025 ↑3 9
- 8
Qwen3-32B ↑9 9
- 9
Gemini 2.0 Flash Feb 2025 9
- 10 GPT-4.1 mini ↓2 9
humanities score mean
35 models we track have a score here.
- 1 gpt-oss-120b ↑6 9
- 2 Claude 3.7 Sonnet ↑3 9
- 3 GPT-5 mini ↑3 9
- 4
Gemini 2.5 Pro Jun 2025 ↓2 9
- 5 GPT-4.1 ↓1 9
- 6 o3 ↓5 9
- 7 GPT-4o Mar 2025 ↑3 9
- 8
Qwen3-32B ↑9 9
- 9
Gemini 2.0 Flash Feb 2025 9
- 10 GPT-4.1 mini ↓2 9
persian general knowledge score w mean
35 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑1 9
- 2
Gemini 2.5 Flash Sep 2025 ↑1 9
- 3 o3 ↓2 8
- 4
Gemini 2.0 Flash Feb 2025 ↑5 8
- 5 Claude 3.7 Sonnet 8
- 6 GPT-4o Mar 2025 ↑4 8
- 7 GPT-4.1 ↓3 8
- 8
Gemini 2.0 Flash-Lite ↑4 8
- 9
Cohere Command A ↑4 6
- 10 GPT-5 mini ↓4 6
persian general knowledge score mean
35 models we track have a score here.
- 1
Gemini 2.5 Pro Jun 2025 ↑1 9
- 2
Gemini 2.5 Flash Sep 2025 ↑1 9
- 3 o3 ↓2 9
- 4
Gemini 2.0 Flash Feb 2025 ↑5 8
- 5 Claude 3.7 Sonnet 8
- 6 GPT-4.1 ↓2 8
- 7 GPT-4o Mar 2025 ↑3 8
- 8
Gemini 2.0 Flash-Lite ↑4 8
- 9
Cohere Command A ↑4 6
- 10 GPT-5 mini ↓4 6
chatbot rag score w mean
35 models we track have a score here.
- 1 o3 10
- 2 gpt-oss-120b ↑5 10
- 3
Mistral Small 3.1 ↑20 10
- 4 GPT-4.1 mini ↑4 9
- 5 gpt-oss-20b ↑9 9
- 6
Qwen3-14B ↑18 9
- 7
Aya Expanse 32B ↑19 9
- 8 Claude 3.7 Sonnet ↓3 9
- 9
Command R+ ↑18 9
- 10
Cohere Command A ↑3 9
chatbot rag score mean
35 models we track have a score here.
- 1 o3 10
- 2 gpt-oss-120b ↑5 10
- 3
Mistral Small 3.1 ↑20 10
- 4 GPT-4.1 mini ↑4 9
- 5
Qwen3-14B ↑19 9
- 6 gpt-oss-20b ↑8 9
- 7 Claude 3.7 Sonnet ↓2 9
- 8
Aya Expanse 32B ↑18 9
- 9 GPT-5 mini ↓3 9
- 10
Command R+ ↑17 9
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official Persian MT-Bench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| o3 | Thinking | 9 | |||
| Thinking | 9 | ||||
| Thinking | 9 | ||||
| GPT-4.1 | Non-thinking | 9 | |||
| Claude 3.7 Sonnet | Non-thinking | 9 | |||
| GPT-5 mini | Thinking | 9 | |||
| gpt-oss-120b | Thinking | 8 | |||
| GPT-4.1 mini | Non-thinking | 8 | |||
| Non-thinking | 8 | ||||
| GPT-4o Mar 2025 | Non-thinking | 8 | |||
| GPT-5 mini | minimal reasoning | 8 | |||
| GPT-5 nano | Thinking | 8 | |||
| Non-thinking | 8 | ||||
| Non-thinking | 8 | ||||
| gpt-oss-20b | Thinking | 8 | |||
| Non-thinking | 8 | ||||
| gpt-oss-120b | low reasoning | 8 | |||
| GPT-4o mini | Non-thinking | 8 | |||
| Non-thinking | 8 | ||||
| Claude 3.5 Haiku | Non-thinking | 8 | |||
| Non-thinking | 8 | ||||
| GPT-5 nano | minimal reasoning | 7 | |||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| gpt-oss-20b | low reasoning | 7 | |||
| GPT-4.1 nano | Non-thinking | 7 | |||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| deepseek-r1-0528-qwen3-8b | Thinking | 7 | |||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 6 | ||||
| Non-thinking | 6 | ||||
| Non-thinking | 4 | ||||
| Non-thinking | 2 |
Scores at each effort level
Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.
Source: Official Persian MT-Bench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| o3 | Thinking | 9 | |||
| Thinking | 9 | ||||
| Thinking | 9 | ||||
| GPT-4.1 | Non-thinking | 9 | |||
| Claude 3.7 Sonnet | Non-thinking | 9 | |||
| GPT-5 mini | Thinking | 9 | |||
| gpt-oss-120b | Thinking | 8 | |||
| GPT-4.1 mini | Non-thinking | 8 | |||
| Non-thinking | 8 | ||||
| GPT-4o Mar 2025 | Non-thinking | 8 | |||
| GPT-5 mini | minimal reasoning | 8 | |||
| GPT-5 nano | Thinking | 8 | |||
| Non-thinking | 8 | ||||
| Non-thinking | 8 | ||||
| gpt-oss-20b | Thinking | 8 | |||
| Non-thinking | 8 | ||||
| gpt-oss-120b | low reasoning | 8 | |||
| GPT-4o mini | Non-thinking | 8 | |||
| Non-thinking | 8 | ||||
| Claude 3.5 Haiku | Non-thinking | 8 | |||
| Non-thinking | 8 | ||||
| GPT-5 nano | minimal reasoning | 7 | |||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| gpt-oss-20b | low reasoning | 7 | |||
| GPT-4.1 nano | Non-thinking | 7 | |||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| deepseek-r1-0528-qwen3-8b | Thinking | 7 | |||
| Non-thinking | 7 | ||||
| Non-thinking | 7 | ||||
| Non-thinking | 6 | ||||
| Non-thinking | 6 | ||||
| Non-thinking | 4 | ||||
| Non-thinking | 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 MT-Bench 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 Persian MT-Bench 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
7 of 9 checks pass, 2 warnings. Look first at the Confidence check. Likely. Some facts rest on secondary sources. Every benchmark
Catalog record 1 warning 1 pass
- Pass Task access All tasks are public.
- Warning Confidence Likely. Some facts rest on secondary sources.
Scores 1 warning 2 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (41 entries).
- Pass Models scored 35 models we track, plus 6 other entries on its leaderboard.
- Warning Newest score The newest came 14 months ago, in September 2025, for Gemini 2.5 Flash.
- Note Frontier models None of the top 10 has a score. The strongest tested is o3 (#84).
- Note Headroom Scored as a rating, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on mcinext-mizan-llm-leaderboard.hf.space: 39 rows for models we track and 6 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.