MTEB: Massive Text Embedding Benchmark

TL;DR

MTEB is a comprehensive benchmark that evaluates 33 text embedding models across 8 tasks and 58 datasets, revealing no single method excels universally, highlighting the need for further research.

Contribution

Introduces the largest and most diverse text embedding benchmark (MTEB) covering multiple tasks, datasets, and languages, with open-source tools and a public leaderboard.

Findings

No embedding method dominates across all tasks

The field has not converged on a universal embedding method

Benchmark covers 58 datasets and 112 languages

Abstract

Text embeddings are commonly evaluated on a small set of datasets from a single task not covering their possible applications to other tasks. It is unclear whether state-of-the-art embeddings on semantic textual similarity (STS) can be equally well applied to other tasks like clustering or reranking. This makes progress in the field difficult to track, as various models are constantly being proposed without proper evaluation. To solve this problem, we introduce the Massive Text Embedding Benchmark (MTEB). MTEB spans 8 embedding tasks covering a total of 58 datasets and 112 languages. Through the benchmarking of 33 models on MTEB, we establish the most comprehensive benchmark of text embeddings to date. We find that no particular text embedding method dominates across all tasks. This suggests that the field has yet to converge on a universal text embedding method and scale it up sufficiently to provide state-of-the-art results on all embedding tasks. MTEB comes with open-source code and a public leaderboard at https://github.com/embeddings-benchmark/mteb.