aavetis/PRarena

This repo tracks the opened and merged PRs by the top SWE coding agents by OpenAI, GitHub, and others. Updates regularly.

What it solves

PRarena provides a way to compare the performance of different AI coding agents by tracking their pull request (PR) volume and success rates on GitHub. It addresses the gap in understanding which AI agents produce the most mergeable code across different development workflows.

How it works

The project uses GitHub search queries to track the number of All PRs, Ready PRs, and Merged PRs created by specific AI agents (such as Copilot, Codex, Cursor, Devin, Codegen, and Jules). By focusing on "Ready PRs" to normalize for agents that use draft PRs differently, it calculates a success rate based on the ratio of merged PRs to ready PRs.

Who it’s for

Developers and researchers who want to benchmark AI coding agents based on real-world GitHub activity and merge success rates.

Highlights

  • own interactive dashboard for visualizing statistics.
  • Tracks multiple major AI agents including Copilot, Codex, Cursor, Devin, and Jules.
  • Normalizes data by distinguishing between draft and ready PRs to ensure fair comparison across different agent workflows.

Related

  • Project
  • Project
  • Project
  • Project
  • Project