Collection · 4 entries

Embedding models for retrieval

Sentence and document embedding models used to build retrieval systems. Dimension and sequence length matter more than size here, and both come from the publisher.

Edited by SAVRN research · last revised 2026-09-18

Entries in this collection

Model · Sentence similarity

all-MiniLM-L6-v2

Published by Sentence Transformers

Why it is here: The small baseline most retrieval stacks start with.

Open weights apache-2.0 23M parameters

View model

Model · Sentence similarity

bge-m3

Published by Beijing Academy of Artificial Intelligence

Why it is here: Multilingual, long-sequence retrieval.

Open weights mit

View model

Model · Feature extraction

multilingual-e5-large

Published by Liang Wang

Why it is here: A widely used multilingual encoder.

Open weights mit 560M parameters

View model

Model · Sentence similarity

all-mpnet-base-v2

Published by Sentence Transformers

Why it is here: A stronger English baseline than MiniLM.

Open weights apache-2.0 109M parameters

View model

Compare the first 4 models in this collection

Explore More

Every collection · The model directory · How a collection is decided