Datasets:
File size: 4,356 Bytes
cb0c27a d3ed3db cb0c27a d3ed3db b4444e8 d3ed3db b4444e8 d3ed3db | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 | ---
license: cc-by-4.0
language:
- ur
pretty_name: UrduMMLU
size_categories:
- 10K<n<100K
task_categories:
- question-answering
- multiple-choice
tags:
- urdu
- mmlu
- multiple-choice
- education
- benchmark
- low-resource
configs:
- config_name: default
data_files:
- split: test
path: urdummlu.json
---
<p align="center">
<img src="logo.png" alt="UrduMMLU" width="180">
</p>
# UrduMMLU
**UrduMMLU** is a large-scale, human-curated benchmark of **26,431** multiple-choice
questions written natively in Urdu. Questions are drawn from Pakistani secondary and
higher-secondary curricula (SSC-I through HSSC-II) and span the humanities, social
sciences, STEM, professional studies, and general knowledge.
Unlike machine-translated MMLU variants, every item here is sourced from native Urdu
exam material, then cleaned, de-duplicated, schema-normalized, and human-verified
through a multi-stage annotation pipeline.
## Dataset at a glance
| | |
|---|---|
| Questions | 26,431 |
| Language | Urdu (`ur`) |
| Format | Single-answer multiple choice (4–5 options) |
| Levels | SSC-I, SSC-II, HSSC-I, HSSC-II |
| Domains | 5 (26 subdomains) |
| Files | `urdummlu.json`, `stats.json` |
## Schema
Each record in `urdummlu.json` has exactly these fields:
```json
{
"id": 0,
"question": "مقبرہ شاہ رکن عالمؒ ____ میں ہے۔",
"options": { "A": "لاہور", "B": "ساہیوال", "C": "سیالکوٹ", "D": "ملتان" },
"correct_key": "D",
"domain": "Social Sciences",
"subdomain": "pakistan studies",
"level": "SSC-II",
"length_tier": "short",
"source": [
{ "name": "BISE Multan 2025", "url": "https://www.bisemultan.edu.pk" }
]
}
```
| Field | Description |
|---|---|
| `id` | Stable integer identifier |
| `question` | Question stem in Urdu |
| `options` | Map of option key (`A`–`E`) to Urdu answer text |
| `correct_key` | Key of the correct option |
| `domain` | Top-level domain (one of 5) |
| `subdomain` | Fine-grained subject (26 total) |
| `level` | Curriculum level: `SSC-I`, `SSC-II`, `HSSC-I`, `HSSC-II` |
| `length_tier` | `short`, `long`, or `null` (question-length bucket) |
| `source` | List of `{name, url}` provenance entries |
## Domain distribution
| Domain | Questions |
|---|---|
| Humanities | 11,010 |
| Social Sciences | 7,968 |
| STEM | 5,113 |
| Other | 1,365 |
| Profession | 975 |
The 26 subdomains include Urdu literature, Urdu language, Islamic studies,
Pakistan studies, chemistry, biology, mathematics, computer science, economics,
sociology, and more. Per-subdomain, per-level, and source counts are in
[`stats.json`](stats.json).
## Curriculum levels
| Level | Questions |
|---|---|
| SSC-I | 11,601 |
| SSC-II | 6,838 |
| HSSC-II | 4,125 |
| HSSC-I | 3,867 |
## Usage
```python
from datasets import load_dataset
ds = load_dataset("MBZUAI/UrduMMLU", split="test")
ex = ds[0]
print(ex["question"])
print(ex["options"], "→", ex["correct_key"])
```
### Evaluation
Treat each item as a single-answer MCQ: present `question` and `options`, then
compare the model's chosen key against `correct_key`. Exact-match accuracy is the
primary metric. We recommend reporting accuracy broken down by `domain` and `level`.
## Sources
Questions were collected from publicly available Urdu exam and practice-question
repositories, including Ustad 360, MCQ Times, TestPoint PK, eTest, FBISE,
ExamAunty, GoTest, PakMCQs, and provincial examination boards (e.g. BISE Multan).
Each item retains its source attribution in the `source` field.
## Limitations
- Answer keys reflect the original source material and the annotation process;
rare errors may remain.
- `length_tier` is `null` for items where a length bucket was not assigned.
- Coverage is weighted toward humanities and the SSC levels, mirroring the
availability of native Urdu exam content.
## Citation
```bibtex
@article{tabassum2026urdummlu,
title = {{UrduMMLU}: A Massive Multitask Benchmark for {Urdu} Language Understanding},
author = {Tabassum, Ahmer and Ahmad, Sarfraz and Iqbal, Hasan and
Aijaz, Owais and Ahsan, Momina and Nakov, Preslav},
journal = {arXiv preprint},
year = {2026},
url = {https://anonymous.for.review}
}
```
## License
Released under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/).
|