MS Marco TinyBERT L2 v2
cross-encoder/ms-marco-TinyBERT-L2-v2
published Mar 2022 · updated Aug 2025
MS Marco TinyBERT L2 v2 is a cross-encoder rerank model trained on the MS MARCO Passage Ranking dataset for information retrieval passage reranking.
specs
| Task | Cross-Encoder Reranking |
| Architecture | TinyBERT with 2 layers |
| Max Sequence Length | 512 tokens |
| Training Data | MS MARCO Passage Ranking |
| NDCG@10 (TREC DL 19) | 69.84 |
about this model
cross-encoder/ms-marco-TinyBERT-L2-v2 is a cross-encoder reranking model optimized for passage ranking, trained on the MS MARCO Passage Ranking dataset. It takes a query and a passage as input and outputs a relevance score, making it suitable for retrieve-and-rerank pipelines where a fast first-stage retriever (e.g., ElasticSearch) returns candidate passages and this model reorders them by relevance.
Key Strengths
- High throughput: processes approximately 9,000 documents per second on a V100 GPU, offering a strong speed-accuracy trade-off.
- Competitive ranking quality on standard benchmarks.
- Lightweight architecture (TinyBERT with 2 layers) designed for low-latency production use.
Performance
The table below reports NDCG@10 on the TREC Deep Learning 2019 task, MRR@10 on the MS Marco Dev set, and inference throughput (docs/sec on V100). The model is compared against several other cross-encoders in the same family.
| Model-Name | NDCG@10 (TREC DL 19) | MRR@10 (MS Marco Dev) | Docs / Sec |
|---|---|---|---|
| cross-encoder/ms-marco-TinyBERT-L2-v2 | 69.84 | 32.56 | 9000 |
| cross-encoder/ms-marco-MiniLM-L2-v2 | 71.01 | 34.85 | 4100 |
| cross-encoder/ms-marco-MiniLM-L4-v2 | 73.04 | 37.70 | 2500 |
| cross-encoder/ms-marco-MiniLM-L6-v2 | 74.30 | 39.01 | 1800 |
| cross-encoder/ms-marco-MiniLM-L12-v2 | 74.31 | 39.02 | 960 |
| cross-encoder/ms-marco-TinyBERT-L2 | 67.43 | 30.15 | 9000 |
| cross-encoder/ms-marco-TinyBERT-L4 | 68.09 | 34.50 | 2900 |
| cross-encoder/ms-marco-TinyBERT-L6 | 69.57 | 36.13 | 680 |
| cross-encoder/ms-marco-electra-base | 71.99 | 36.41 | 340 |
| nboost/pt-tinybert-msmarco | 63.63 | 28.80 | 2900 |
| nboost/pt-bert-base-uncased-msmarco | 70.94 | 34.75 | 340 |
| nboost/pt-bert-large-msmarco | 73.36 | 36.48 | 100 |
| Capreolus/electra-base-msmarco | 71.23 | 36.89 | 340 |
| amberoad/bert-multilingual-passage-reranking-msmarco | 68.40 | 35.54 | 330 |
| sebastian-hofstaetter/distilbert-cat-margin_mse-T2-msmarco | 72.82 | 37.88 | 720 |
This model is hosted by gigarouter as a managed, OpenAI‑compatible API. Developers can call it directly without managing infrastructure or model weights.
We're benchmarking and onboarding MS Marco TinyBERT L2 v2 as a hosted, OpenAI-compatible API. Sign in for free credit and be ready when it lands, or tell us you want it and we'll prioritize it.