MTEB logo MTEB
Persian icon

Persian

Cite this benchmark

@article{zinvandi2025famteb,
  author = {Zinvandi, Erfan and Alikhani, Morteza and Sarmadi, Mehran and Pourbahman, Zahra and Arvin, Sepehr and Kazemi, Reza and Amini, Arash},
  journal = {arXiv preprint arXiv:2502.11571},
  title = {Famteb: Massive text embedding benchmark in persian language},
  year = {2025},
}

Persian text embedding quality across classification, clustering, pair classification, reranking, retrieval, semantic similarity, and summarization retrieval. In v2, large datasets were optimized for accessibility, low-quality datasets were removed, and higher-quality data was added; see the main PR for details.

Reference paper →
Languages 1
Tasks 52
Task Types 7
Models 0