According to a Hugging Face blog post, the organization has introduced the Ettin reranker family. The headline frames this as a new addition to the ecosystem of models for information retrieval. Rerankers typically work after a first-stage retrieval step, re-scoring candidate documents to improve the final ordering.
The source is a single announcement, so the available information is limited to the fact of the release. No benchmarks, model sizes, or usage instructions are provided in the headline. Readers looking for technical specifics would need to consult the full blog post for details.