Find an embedding model

Vector models for retrieval, RAG, similarity, clustering, and deduplication. Bearing ranks against MTEB, pricing, multilingual coverage, and self-hosting options.

What's it for?

Retrieval / RAGselected

Find relevant docs for a query (most common)

Semantic similarity

Score how similar two pieces of text are

Classification

Use embeddings as input features for a classifier

Clustering

Group similar items without labels

Deduplication

Find near-duplicate records at scale

Other / not sure

Pick this and we will rank by general MTEB quality

How long are the texts you'll embed?

Short

Queries, sentences, social posts

Mediumselected

Paragraphs, FAQ entries, abstracts

Long

Full documents, contracts, chapters

Hosting preference

No preferenceselected

Hosted APIs and open models both eligible

Prefer hosted API

Lowest operational burden

Prefer open / self-hosted

Full control + no data egress

Languages

English onlyselected

English-optimised models score higher on MTEB-Eng

English plus a few others

European + a handful of other languages

Many languages

20+ languages — surfaces multilingual specialists

Latency

Batch — latency doesn’t matter

Run overnight to build / refresh an index

Interactiveselected

Embed a user query at request time

Realtime

Sub-100 ms per query, hot paths

Built by The Good Ship · tomcw.xyz

MIT License