What the study found
This systematic literature review maps how large language models, or LLMs, have been used in automated program repair from 2020 to 2025. The authors report that they analyzed 189 relevant papers and organized the literature around LLM types, deployment strategies, repair scenarios, and research challenges.
Why the authors say this matters
The authors conclude that their review gives the automated program repair community a systematic overview of the research landscape. They say it can help researchers understand current achievements, challenges, and opportunities, and support future research.
What the researchers tested
The paper is a systematic literature review of LLM-based automated program repair work published between 2020 and 2025. The authors analyzed 189 relevant papers from the perspectives of large language models, automated program repair, and how the two are combined.
What worked and what didn't
The review found that existing papers use a range of popular LLMs to support automated program repair and that the authors identified four utilization strategies for deployment. It also describes repair scenarios that benefit from LLMs, including semantic bugs and security vulnerabilities, and notes several integration issues such as input forms and open science. The abstract also says the paper highlights remaining challenges and potential guidelines for future research.
What to keep in mind
The abstract does not provide the detailed results for each category or the specific challenges and guidelines mentioned. It also does not report quantitative comparisons among techniques, so the level of effectiveness of individual approaches is not described in the available summary.
Key points
- The paper is a systematic literature review of large language models in automated program repair.
- It covers research published between 2020 and 2025 and includes 189 relevant papers.
- The authors categorized popular LLMs, four deployment strategies, and repair scenarios such as semantic bugs and security vulnerabilities.
- The review also discusses integration issues, including input forms and open science.
- The abstract says the paper highlights remaining challenges and potential guidelines for future research.
Disclosure
- Research title:
- Review maps large language models in automated program repair
- Publication date:
- 2026-03-02
- DOI:
- 10.1145/3799693
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.