{
    "Name": "Mizan",
    "Dialect Subsets": [
        {
            "Name": "Iraqi Track",
            "Volume": 265.0,
            "Unit": "sentences",
            "Dialect": "Iraq"
        },
        {
            "Name": "MSA Track",
            "Volume": 75.0,
            "Unit": "sentences",
            "Dialect": "Modern Standard Arabic"
        }
    ],
    "HF Link": "https://huggingface.co/datasets/nawaralseelawi/mizan-iraqi-arabic-benchmark",
    "Link": "https://huggingface.co/datasets/nawaralseelawi/mizan-iraqi-arabic-benchmark",
    "License": "CC BY 4.0",
    "Year": 2026,
    "Language": "ar",
    "Dialect": "mixed",
    "Source": [
        "manual construction"
    ],
    "Domain": [
        "politics",
        "culture",
        "society",
        "geography",
        "history",
        "law"
    ],
    "Form": "text",
    "Annotation Style": [
        "human annotation",
        "human validation"
    ],
    "Description": "National benchmark for evaluating LLMs on Iraqi Arabic and civic context.",
    "Volume": 340.0,
    "Unit": "sentences",
    "Provider": [
        "University of Misan"
    ],
    "Derived From": [],
    "Paper Title": "Mizan: A National Benchmark for Evaluating Large Language Models on Iraqi Arabic and the Iraqi Civic Context",
    "Paper Link": "https://arxiv.org/pdf/2609.13980.pdf",
    "Script": "Arab",
    "Tokenized": false,
    "Host": "other",
    "Access": "Free",
    "Cost": "",
    "Has Splits": false,
    "Partial": false,
    "Tasks": [
        "question answering",
        "multiple choice question answering",
        "machine translation",
        "information extraction"
    ],
    "Venue Title": "arXiv",
    "Venue Type": "preprint",
    "Venue Name": "arXiv",
    "Authors": [
        "Nawar S. Alseelawi",
        "Mustafa S. Aljumaily"
    ],
    "Affiliations": [
        "University of Misan",
        "Missan Oil Company"
    ],
    "Abstract": "Arabic large-language-model (LLM) evaluation has matured around Modern Standard Arabic (MSA). Dialectal Arabic - the language Iraqis actually speak - remains nearly invisible to this infrastructure. We introduce Mizan (\u201dthe balance\u201d), Iraq\u2019s national benchmark for evaluating LLMs on Iraqi Arabic and the Iraqi civic context: an MSA baseline track paired with an Iraqi track across six axes - dialect comprehension, dialect generation, bidirectional MSA-Iraqi translation, Iraq-specific knowledge, official-document field extraction, and safety - built from 340 originally authored, dually reviewed items with statistically audited answer positions and Wilson intervals. A pilot evaluation of 27 systems yields four findings. The MSA track saturates while the Iraqi track discriminates. Official-document extraction confines every system to 32-56. Arabic-focused specialization behaves as MSA specialization. The safety-hardened tier of the newest model family deterministically refuses innocuous dialect-comprehension items as policy violations.",
    "Added By": "Zaid Alyafeai"
}