Agentic Much? Adoption of Coding Agents on GitHub

TL;DR

This study provides the first large-scale, artifact-based analysis of coding agent adoption on GitHub, revealing rapid, broad uptake from early 2025 onward and documenting adoption rates between 15.85% and 22.60% depending on estimation approach. By leveraging explicit traces left in files, commits, and PRs, the authors quantify adoption at both file and commit levels and characterize how adoption correlates with project size, age, and activity. They show that AI-assisted commits tend to be larger and more feature- or bug-focused than human commits, indicating a departure from traditional code-completion tools toward autonomous, substantive contributions. The findings underscore the practical and research significance of coding agents, highlighting the need for careful attribution, ongoing study of their impact on productivity and code quality, and broader understanding of how these agents reshape software development workflows.

Abstract

In the first half of 2025, coding agents have emerged as a category of development tools that have very quickly transitioned to the practice. Unlike ''traditional'' code completion LLMs such as Copilot, agents like Cursor, Claude Code, or Codex operate with high degrees of autonomy, up to generating complete pull requests starting from a developer-provided task description. This new mode of operation is poised to change the landscape in an even larger way than code completion LLMs did, making the need to study their impact critical. Also, unlike traditional LLMs, coding agents tend to leave more explicit traces in software engineering artifacts, such as co-authoring commits or pull requests. We leverage these traces to present the first large-scale study (129,134 projects) of the adoption of coding agents on GitHub, finding an estimated adoption rate of 15.85%--22.60%, which is very high for a technology only a few months old--and increasing. We carry out an in-depth study of the adopters we identified, finding that adoption is broad: it spans the entire spectrum of project maturity; it includes established organizations; and it concerns diverse programming languages or project topics. At the commit level, we find that commits assisted by coding agents are larger than commits only authored by human developers, and have a large proportion of features and bug fixes. These findings highlight the need for further investigation into the practical use of coding agents.

Figures (12)

• Figure 1: Typical workflow of a coding agent
• Figure 2: Evolution of the number of AI coding agents since 2022
• Figure 3: Pairwise relations of file and commit level adoption metrics
• Figure 4: Topics of GitHub projects showing file-based evidence of the adoption of coding agents, for topics associated to at least 350 projects (with file-based evidence of use). Each topic is plotted with adoption rate, i.e., ratio of projects using agents among projects with the same topic, on the X axis; and the number of projects using agents on the Y axis (log scale).
• Figure 5: Commit Ratio distribution by topic of commit-level adopters projects