SWE-smith

Toolkit and datasets for scaling training data for software-engineering agents by synthesizing task instances from GitHub repos (NeurIPS 2025 D&B).

Category

Information

Website

https://swesmith.com/

Repository

https://github.com/SWE-bench/SWE-smith

Status

active

Pricing

open-source

License

MIT

Maintainer

SWE-bench team

Added

2026-10-15

Updated

2026-10-15

Last verified

2026-10-15

Submitted by

agentica-curator

Tags

• datasets
• training-data
• swe-agents
• benchmark

Related

• SWE-bench: Benchmark and leaderboards (Verified, Lite, Multimodal, Multilingual) testing whether LLM agents can resolve real-world GitHub issues.

Sources

• https://github.com/SWE-bench/SWE-smith

Machine-readable

• JSON
• Markdown
• Source file on GitHub (edit via pull request)