{"ID":2854755,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.14880","arxiv_id":"2510.14880","title":"Fantastic (small) Retrievers and How to Train Them: mxbai-edge-colbert-v0 Tech Report","abstract":"In this work, we introduce mxbai-edge-colbert-v0 models, at two different parameter counts: 17M and 32M. As part of our research, we conduct numerous experiments to improve retrieval and late-interaction models, which we intend to distill into smaller models as proof-of-concepts. Our ultimate aim is to support retrieval at all scales, from large-scale retrieval which lives in the cloud to models that can run locally, on any device. mxbai-edge-colbert-v0 is a model that we hope will serve as a solid foundation backbone for all future experiments, representing the first version of a long series of small proof-of-concepts. As part of the development of mxbai-edge-colbert-v0, we conducted multiple ablation studies, of which we report the results. In terms of downstream performance, mxbai-edge-colbert-v0 is a particularly capable small model, outperforming ColBERTv2 on common short-text benchmarks (BEIR) and representing a large step forward in long-context tasks, with unprecedented efficiency.","short_abstract":"In this work, we introduce mxbai-edge-colbert-v0 models, at two different parameter counts: 17M and 32M. As part of our research, we conduct numerous experiments to improve retrieval and late-interaction models, which we intend to distill into smaller models as proof-of-concepts. Our ultimate aim is to support retrieva...","url_abs":"https://arxiv.org/abs/2510.14880","url_pdf":"https://arxiv.org/pdf/2510.14880v1","authors":"[\"Rikiya Takehi\",\"Benjamin Clavié\",\"Sean Lee\",\"Aamir Shakir\"]","published":"2025-10-16T17:00:35Z","proceeding":"cs.IR","tasks":"[\"cs.IR\"]","methods":"[]","has_code":false}