We introduce control models for LLM-powered code completion in JetBrains IDEs: ML classifiers which trigger inference and filter the generated suggestions to better align them with users and reduce unnecessary requests. To this end, we evaluate boosting- and transformer-based architectures on an offline dataset of real code completions with n=98 users. We further evaluate the offline classification performance of our boosting-based approach on a range of syntactically diverse languages; and perform an A/B study in a production environment where they improve completion efficiency and quality metrics. With this study, we hope to demonstrate the potential in using auxiliary models for smarter in-IDE integration of LLM-driven features, highlight fruitful future directions, and open problems.
@article{arxiv.2601.20223,
title = {Control Models for In-IDE Code Completion},
author = {Aral de Moor and Yana Hrynevich and Hleb Badzeika and Vladyslav Furda and Marko Kojic and Artem Savelev and Kostadin Cvejoski and Darya Rovdo and Ekaterina Garanina},
journal= {arXiv preprint arXiv:2601.20223},
year = {2026}
}
Comments
6 pages; accepted at IDE'26 co-located with ICSE'26