MS MARCO Cross-Encoders
UKP Lab (TU Darmstadt)The `cross-encoder/ms-marco-*` family of BERT and MiniLM cross-encoders trained on MS MARCO Passage Ranking and published alongside SentenceTransformers. Each checkpoint takes a (query, passage) pair and returns a single relevance logit, and the family spans TinyBERT-L2 through MiniLM-L12 so a deployment can trade throughput against NDCG. It is the default self-hosted reranker in most retrieve-and-rerank stacks.
Added 2026-09-11 for the embeddings_retrieval promotion. The family is old - the v2 checkpoints date to 2022 - and English-only, but it remains by a wide margin the most downloaded reranker line on the Hub. Verified 2026-09-11 via the ms-marco-MiniLM-L6-v2 model card and the SentenceTransformers MS MARCO cross-encoder training directory.
Openness
5 high confidence- weights
- open(Apache-2.0 ungated safetensors for every checkpoint in the family)
- data
- open(MS MARCO Passage Ranking
- code
- open(SentenceTransformers ships the MS MARCO cross-encoder training scripts the card links to)
- license
- Apache-2.0(OSI)
The rare case where the whole chain is public: an Apache-2.0 ungated checkpoint, a public training corpus named in the card front matter, and the actual training scripts that produced it shipped in the SentenceTransformers examples tree. The ladder reaches its top rung on the evidence rather than on reputation.
- https://huggingface.co/cross-encoder/ms-marco-MiniLM-L6-v2/raw/main/README.md recorded 2026-09-11
Front matter reads license: apache-2.0 and datasets: sentence-transformers/msmarco; the body says the model was trained on MS MARCO Passage Ranking and links the SBERT.net MS MARCO training code.
- https://raw.githubusercontent.com/UKPLab/sentence-transformers/master/examples/cross_encoder/training/ms_marco/README.md recorded 2026-09-11
The cross_encoder/training/ms_marco directory of SentenceTransformers, listing training_ms_marco_bce.py, training_ms_marco_cmnrl.py, training_ms_marco_listnet.py and training_ms_marco_lambda.py as the scripts that train a CrossEncoder on MS MARCO.
Adoption
5 high confidenceAbout 96.5M downloads in the trailing 30 days summed across the cross-encoder/ms-marco-* checkpoints, of which MiniLM-L6-v2 alone is 87.5M and MiniLM-L4-v2 a further 5.7M. Either figure on its own clears the >10M band; the family total is the largest of any reranker on the map.
- https://huggingface.co/api/models/cross-encoder/ms-marco-MiniLM-L6-v2 recorded 2026-09-11
downloads: 87478105; gated: false; private: false; license apache-2.0 for cross-encoder/ms-marco-MiniLM-L6-v2.
- https://huggingface.co/api/models/cross-encoder/ms-marco-MiniLM-L4-v2 recorded 2026-09-11
downloads: 5701227; gated: false; private: false for cross-encoder/ms-marco-MiniLM-L4-v2.
Capability
2 high confidenceRung 2, narrow or dated, read straight off the category definition: the v2 checkpoints date to 2022, take English only, and cap at 512 tokens. The published numbers are credible and the family is still the default self-hosted reranker, but a rung is a statement about retrieval quality and scope rather than about popularity. One rung below bge-reranker, which is multilingual, longer-context and a generation newer.
- https://huggingface.co/cross-encoder/ms-marco-MiniLM-L6-v2/raw/main/README.md recorded 2026-09-11
Performance table gives cross-encoder/ms-marco-MiniLM-L6-v2 at 74.30 NDCG@10 on TREC DL 19 and 39.01 MRR@10 on MS MARCO dev, with the family spanning TinyBERT-L2 at 69.84 to MiniLM-L12 at 74.31; language front matter is en.
Verified 2026-09-11