Journal article · 2026 · Open access
Retrofitting language models to operate over bytes
Nature, 658(8137), 915–922 · Published 7 October 2026 · DOI 10.1038/s41586-026-11111-4
In brief
“Retrofitting language models to operate over bytes” is a 2026 journal article by Benjamin Minixhofer and 8 co-authors, published in Nature. OpenAlex counts 22 citations as of 16 November 2026, which puts it in the top 1% of works published in 2026 by citations. It is open access under a CC BY licence: the full text is free to read at Nature.
- Citations
- 22
- Most cited in
- 2026 (22)
- Field-weighted impact
- 30.73
- References
- 89
- Authors
- 9
Abstract
Recent advances in artificial intelligence (AI) have largely been driven by large language models, deep neural networks that operate over discrete units called tokens. To represent text, most large language models use words or word fragments as the tokens, known as subword tokenization 1 . Subword tokenization obscures fine-grained information, which is problematic, especially for scientific data—such as computer code or biological sequences—where meaning depends on the individual characters or bytes 2 . Models that instead operate directly on the byte encoding of text avoid these limitations, but until now they have lagged behind subword-based models in performance. Here we introduce a general method for creating byte-level large language models through byteification that approach the capabilities of subword-based systems. We use a two-stage conversion procedure to retrofit existing subword-based models into byte-level models with minimal extra training. The resulting models outperform earlier byte-level approaches and excel on character-level reasoning tasks, achieving practical inference speeds by efficiently processing byte-level information and adaptability by reusing the existing ecosystem around the source large language model. Our results remove a long-standing performance barrier to end-to-end byte-level language modelling, demonstrating that models operating on raw text encodings can scale competitively while offering advantages in domains requiring fine-grained textual understanding.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Minixhofer, B., Murray, T., Limisiewicz, T., Korhonen, A., Zettlemoyer, L., Smith, N. A., De Ponti, E., Soldaini, L., & Hofmann, V. (2026). Retrofitting language models to operate over bytes. Nature, 658(8137), 915–922. https://doi.org/10.1038/s41586-026-11111-4
MLA 9
Minixhofer, Benjamin, et al. “Retrofitting language models to operate over bytes.” Nature, vol. 658, no. 8137, 2026, pp. 915–922. https://doi.org/10.1038/s41586-026-11111-4.
Chicago
Minixhofer, Benjamin, Tyler Murray, Tomasz Limisiewicz, Anna Korhonen, Luke Zettlemoyer, Noah A. Smith, E De Ponti, Luca Soldaini, and Valentin Hofmann. 2026. “Retrofitting language models to operate over bytes.” Nature 658 (8137): 915–922. https://doi.org/10.1038/s41586-026-11111-4.
Harvard
Minixhofer, B. et al. (2026) ‘Retrofitting language models to operate over bytes’, Nature, 658(8137), pp. 915–922. doi:10.1038/s41586-026-11111-4.
IEEE
B. Minixhofer et al., “Retrofitting language models to operate over bytes,” Nature, vol. 658, no. 8137, pp. 915–922, 2026, doi: 10.1038/s41586-026-11111-4.
BibTeX
@article{minixhofer2026retrofitting,
title = {{Retrofitting language models to operate over bytes}},
author = {Minixhofer, Benjamin and Murray, Tyler and Limisiewicz, Tomasz and Korhonen, Anna and Zettlemoyer, Luke and Smith, Noah A. and De Ponti, E and Soldaini, Luca and Hofmann, Valentin},
journal = {Nature},
year = {2026},
volume = {658},
number = {8137},
pages = {915--922},
doi = {10.1038/s41586-026-11111-4},
url = {https://doi.org/10.1038/s41586-026-11111-4}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- subword tokenization
- large language models
- inference efficiency
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Minixhofer, B., Murray, T., Limisiewicz, T., Korhonen, A., Zettlemoyer, L., Smith, N. A., De Ponti, E., Soldaini, L., & Hofmann, V. (2026). Retrofitting language models to operate over bytes. Nature, 658(8137), 915–922. https://doi.org/10.1038/s41586-026-11111-4
How many times has this paper been cited?
22 times, according to OpenAlex on 16 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at Nature: sign in and open the PDF from this page.
Who wrote this paper?
Benjamin Minixhofer, Tyler Murray, Tomasz Limisiewicz, Anna Korhonen, Luke Zettlemoyer, Noah A. Smith, E De Ponti, Luca Soldaini, Valentin Hofmann.
What is the DOI of this paper?
10.1038/s41586-026-11111-4, permanently at https://doi.org/10.1038/s41586-026-11111-4. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 16 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.