Mizan - Iraqi Arabic: leaderboard
The Iraqi-dialect track of Mizan, the language Iraqis actually speak, which is where the frozen pilot board's ranking is decided. Auto-scored; higher is better. Source: arxiv.org/abs/2609.13980.
Metric: Score (%). Source: arxiv.org. Saturation forecast: Around February 2027. 27 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 84.9 | #159 |
| 2 | GPT-5 | 84.9 | #105 |
| 3 | GPT-6 Astra | 84.4 | #6 |
| 4 | GPT-4o | 83.7 | #357 |
| 5 | Mistral Large 3 | 83.7 | #410 |
| 6 | Gemini 2.5 Flash | 83.6 | #255 |
| 7 | Claude Sonnet 4.5 | 83.6 | #148 |
| 8 | Claude Sonnet 5 | 83.1 | #65 |
| 9 | Llama 4 Maverick | 83 | #486 |
| 10 | Gemma 3 27B (IT) | 82.7 | #528 |
| 11 | Gemma 4 31B (IT) | 82.4 | #178 |
| 12 | GPT-OSS-120B | 82 | #346 |
| 13 | Claude Opus 5 | 82 | #16 |
| 14 | Claude Haiku 4.5 | 81.9 | #287 |
| 15 | DeepSeek V3.2 | 80.5 | #210 |
No result here: #3 Claude Opus 5.5, #7 GPT-6.1 Sol, #9 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=mizan-iraqi-arabic · How It Works · Data refreshed daily, snapshot 2026-11-05.