SwingArena: Competitive Programming Arena for Long-context GitHub Issue Solving
This provides a scalable evaluation method for LLMs in software development, addressing the need for benchmarks that mirror real-world workflows, though it is incremental as it builds on existing retrieval-augmented and interactive evaluation concepts.
The authors tackled the problem of evaluating Large Language Models (LLMs) in realistic software development by introducing SwingArena, a competitive framework that pairs LLMs as submitters and reviewers to generate patches and test cases through continuous integration pipelines, with experiments on over 400 GitHub issues showing that models like GPT-4o excel at aggressive patch generation while DeepSeek and Gemini prioritize correctness.
We present SwingArena, a competitive evaluation framework for Large Language Models (LLMs) that closely mirrors real-world software development workflows. Unlike traditional static benchmarks, SwingArena models the collaborative process of software iteration by pairing LLMs as submitters, who generate patches, and reviewers, who create test cases and verify the patches through continuous integration (CI) pipelines. To support these interactive evaluations, we introduce a retrieval-augmented code generation (RACG) module that efficiently handles long-context challenges by providing syntactically and semantically relevant code snippets from large codebases, supporting multiple programming languages (C++, Python, Rust, and Go). This enables the framework to scale across diverse tasks and contexts while respecting token limitations. Our experiments, using over 400 high-quality real-world GitHub issues selected from a pool of 2,300 issues, show that models like GPT-4o excel at aggressive patch generation, whereas DeepSeek and Gemini prioritize correctness in CI validation. SwingArena presents a scalable and extensible methodology for evaluating LLMs in realistic, CI-driven software development settings. More details are available on our project page: swing-bench.github.io