SWE-Bench++: A Framework for the Scalable Generation of Software Engineering Benchmarks from Open-Source Repositories

📅 2025-12-19
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
Existing repository-level software engineering benchmarks (e.g., SWE-bench) rely on manual annotation, static datasets, and are limited to Python defect repair, lacking multilingual support and scalability. This work introduces the first scalable, multilingual (11 languages) repository-level evaluation benchmark, automatically constructed from real GitHub Pull Requests to yield executable defect repair and feature development tasks. We propose a novel PR-driven, four-stage automation pipeline: programmatic collection, containerized environment synthesis, test oracle extraction, and quality assurance—augmented by prompt-guided failure trajectory synthesis for model training. The benchmark comprises 11,133 instances across 3,971 repositories. On a 1,782-sample subset, Claude-3.5-Sonnet achieves 36.20% pass@10. Fine-tuning on this benchmark significantly improves performance on SWE-bench Multilingual, demonstrating its efficacy for training and evaluation of multilingual code intelligence models.

Technology Category

Application Category

📝 Abstract
Benchmarks like SWE-bench have standardized the evaluation of Large Language Models (LLMs) on repository-level software engineering tasks. However, these efforts remain limited by manual curation, static datasets, and a focus on Python-based bug fixes. We introduce SWE-Bench++, an automated framework that generates repository-level coding tasks from open-source GitHub projects. Unlike synthetic approaches, our pipeline harvests live pull requests to cover both bug fixes and feature requests across 11 languages. SWE-Bench++ turns GitHub pull requests (PRs) into reproducible, execution-based tasks via four stages: programmatic sourcing, environment synthesis, test oracle extraction, and quality assurance. A final hint-guided trajectory synthesis step converts instances that strong models fail on into training trajectories. Our initial benchmark consists of 11,133 instances from 3,971 repositories across 11 languages. On a subset of 1,782 instances of this benchmark, today's strongest models perform as follows: claude-sonnet-4.5 achieves 36.20% pass@10, gpt-5-2025-08-07 34.57%, gemini/gemini-2.5-pro 24.92%, and gpt-4o 16.89%. We further demonstrate the utility of our dataset by showing that fine-tuning on SWE-Bench++ instances yields measurable improvements on the SWE-bench Multilingual benchmark. SWE-Bench++ provides a scalable, multilingual benchmark for evaluating and improving repository-level code generation.
Problem

Research questions and friction points this paper is trying to address.

Automates generation of software engineering benchmarks from GitHub pull requests
Expands beyond Python bug fixes to 11 languages and feature requests
Creates reproducible tasks for evaluating and training large language models
Innovation

Methods, ideas, or system contributions that make the work stand out.

Automated framework generates coding tasks from GitHub pull requests
Pipeline harvests live pull requests across 11 programming languages
Converts pull requests into reproducible execution-based tasks via four stages
🔎 Similar Papers
No similar papers found.
L
Lilin Wang
Research & Development, Turing
L
Lucas Ramalho
Research & Development, Turing
A
Alan Celestino
Research & Development, Turing
P
Phuc Anthony Pham
Research & Development, Turing
Y
Yu Liu
Research & Development, Turing
U
Umang Kumar Sinha
Research & Development, Turing
A
Andres Portillo
Research & Development, Turing
O
Onassis Osunwa
Research & Development, Turing
G
Gabriel Maduekwe
Research & Development, Turing