English

Cite this benchmark

@article{muennighoff2022mteb,
  author = {Muennighoff, Niklas and Tazi, Nouamane and Magne, Loïc and Reimers, Nils},
  doi = {10.48550/ARXIV.2210.07316},
  journal = {arXiv preprint arXiv:2210.07316},
  publisher = {arXiv},
  title = {MTEB: Massive Text Embedding Benchmark},
  url = {https://arxiv.org/abs/2210.07316},
  year = {2022},
}

English text embedding quality across classification, clustering, retrieval, reranking, pair classification, and semantic similarity. We recommend using MTEB(eng, v2) instead, which resolves a known scoring bug, uses updated task versions, and removes common fine-tuning datasets such as MSMARCO for more comparable scores.

Languages 1

Tasks 56

Task Types 7

Models 0