Introducing the Ettin Reranker series
Original title:Introducing the Ettin Reranker Family
AI summary
Hugging Face released six Ettin Reranker models (17m, 32m, 68m, 150m, 400m, 1b), based on the Ettin ModernBERT encoder and trained by distilling self-mxbai-rerank-large-v2 scores, achieving SOTA at each scale on MTEB(eng, v2) Retrieval. The models are provided via the Sentence Transformers CrossEncoder interface and can be called in three lines of code. Also released is the train-sentence-transformers Agent Skill (v5.5.0), allowing AI coding agents to fine-tune models on user data. The training recipe, datasets, and scripts are all open-sourced.
Key Takeaways
- Hugging Face released six Ettin Reranker models (17m,32m,68m,150m,400m,1b),based on the Ettin…
- The models are provided via the Sentence Transformers CrossEncoder interface and can be called in…
- Also released is the train-sentence-transformers Agent Skill (v5.
- 0),allowing AI coding agents to fine-tune models on user data.
- The training recipe,datasets,and scripts are all open-sourced.
Quick Facts
- Published
- Updated
- Source
- Hugging Face 社区博客(混合发现)
- Category
- Model Releases
- AI score
- 60
从17M到1B的全尺寸reranker家族,每个量级都是SOTA,而且训练数据和代码全开放,做搜索和RAG的开发者可以无痛替换旧模型。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: Introducing the Ettin Reranker series. May 19, 2026. https://new.nat6.net/en/news/cmpcrgqiy04rdslaewcj0tfqm/