Datasets:

MTSAIR
/

MWS-Vision-Bench

pretty_name: MWS Vision Bench
dataset_name: mws-vision-bench
language:
  - ru
license: cc-by-4.0
tags:
  - benchmark
  - multimodal
  - ocr
  - kie
  - grounding
  - vlm
  - business
  - russian
  - document
  - visual-question-answering
  - document-question-answering
task_categories:
  - visual-question-answering
  - document-question-answering
size_categories:
  - 1K<n<10K
annotations_creators:
  - expert-generated
dataset_creators:
  - MTS AI Research
papers:
  - title: >-
      MWS Vision Bench: The First Russian Business-OCR Benchmark for Multimodal
      Models
    authors:
      - MTS AI Research Team
    year: 2025
    status: in preparation
    note: Paper coming soon
homepage: https://huggingface.co/datasets/MTSAIR/MWS-Vision-Bench
repository: https://github.com/mts-ai/MWS-Vision-Bench
organization: MTSAIR

MWS-Vision-Bench

🇷🇺 Русскоязычное описание ниже / Russian summary below.

MWS Vision Bench — the first Russian-language business-OCR benchmark designed for multimodal large language models (MLLMs).
This is the validation split - publicly available for open evaluation and comparison.
🧩 Paper is coming soon.

🔗 Official repository: github.com/mts-ai/MWS-Vision-Bench
🏢 Organization: MTSAIR on Hugging Face
📰 Article on Habr (in Russian): “MWS Vision Bench — the first Russian business-OCR benchmark”

📊 Dataset Statistics

Total samples: 1,302
Unique images: 400
Task types: 5

🖼️ Dataset Preview

Examples of diverse document types in the benchmark: business documents, handwritten notes, technical drawings, receipts, and more.

📁 Repository Structure

MWS-Vision-Bench/
├── metadata.jsonl       # Dataset annotations
├── images/              # Image files organized by category
│   ├── business/
│   │   ├── scans/
│   │   ├── sheets/
│   │   ├── plans/
│   │   └── diagramms/
│   └── personal/
│       ├── hand_documents/
│       ├── hand_notebooks/
│       └── hand_misc/
└── README.md            # This file

📋 Data Format

Each line in metadata.jsonl contains one JSON object:

{
  "file_name": "images/image_0.jpg",   # Path to the image
  "id": "1",                           # Unique identifier
  "type": "text grounding ru",         # Task type
  "dataset_name": "business",          # Subdataset name
  "question": "...",                   # Question in Russian
  "answers": ["398", "65", ...]        # List of valid answers (as strings)
}

🎯 Task Types

Task	Description	Count
`document parsing ru`	Parsing structured documents	243
`full-page OCR ru`	End-to-end OCR on full pages	144
`key information extraction ru`	Extracting key fields	119
`reasoning VQA ru`	Visual reasoning in Russian	400
`text grounding ru`	Text–region alignment	396

📊 Leaderboard (Validation Set)

Top models evaluated on this validation dataset:

Model	Overall	img→text	img→markdown	Grounding	KIE (JSON)	VQA
Gemini-2.5-pro	0.682	0.836	0.745	0.084	0.891	0.853
Gemini-2.5-flash	0.644	0.796	0.683	0.067	0.841	0.833
gpt-4.1-mini	0.643	0.866	0.724	0.091	0.750	0.782
Claude-4.5-Sonnet	0.639	0.723	0.676	0.377	0.728	0.692
gpt-5-mini	0.632	0.797	0.678	0.126	0.784	0.776
Qwen2.5-VL-72B	0.631	0.848	0.712	0.220	0.644	0.732
gpt-5-mini (responses)	0.594	0.743	0.567	0.118	0.811	0.731
Qwen3-VL-30B-A3B	0.589	0.802	0.688	0.053	0.661	0.743
gpt-4.1	0.587	0.709	0.693	0.086	0.662	0.784
Qwen3-VL-32B	0.585	0.732	0.646	0.054	0.724	0.770
Qwen3-VL-30B-A3B-FP8	0.583	0.798	0.683	0.056	0.638	0.740
Qwen2.5-VL-32B	0.577	0.767	0.649	0.232	0.493	0.743
gpt-5 (responses)	0.573	0.746	0.650	0.080	0.687	0.704
Qwen2.5-VL-7B	0.549	0.779	0.704	0.185	0.426	0.651
gpt-4.1-nano	0.503	0.676	0.672	0.028	0.567	0.573
gpt-5-nano	0.503	0.487	0.583	0.091	0.661	0.693
Qwen3-VL-2B	0.439	0.592	0.613	0.029	0.356	0.605
Qwen2.5-VL-3B	0.402	0.613	0.654	0.045	0.203	0.494

Scale: 0.0 - 1.0 (higher is better)

📝 Submit your model: To evaluate on the private test set, contact [email protected]

💻 Usage Example

from datasets import load_dataset

# Load dataset (authorization required if private)
dataset = load_dataset("MTSAIR/MWS-Vision-Bench", token="hf_...")

# Example iteration
for item in dataset:
    print(f"ID: {item['id']}")
    print(f"Type: {item['type']}")
    print(f"Question: {item['question']}")
    print(f"Image: {item['image_path']}")
    print(f"Answers: {item['answers']}")

📄 License

See LICENSE for details.

📚 Citation

If you use this dataset in your research, please cite:

@misc{mwsvisionbench2024,
  title={MWS-Vision-Bench: Russian Multimodal OCR Benchmark},
  author={MTS AI Research},
  organization={MTSAIR},
  year={2025},
  url={https://huggingface.co/datasets/MTSAIR/MWS-Vision-Bench},
  note={Paper coming soon}
}

🤝 Contacts

Team: MTSAIR Research
Email: [email protected]

🇷🇺 Краткое описание

MWS Vision Bench — первый русскоязычный бенчмарк для бизнес-OCR в эпоху мультимодальных моделей.
Он включает 1302 примера и 5 типов задач, отражающих реальные сценарии обработки бизнес-документов и рукописных данных.
Датасет создан для оценки и развития мультимодальных LLM в русскоязычном контексте.
📄 Научная статья в процессе подготовки (paper coming soon).

Made with ❤️ by MTS AI Research Team