Reproduce “RAT: RunAnyThing via Fully Automated Environment Configuration” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Renhong Huang, Dongdong Hua, Yifei Sun, Sitao Ding, Hanyang Yuan, Daixin Wang, Yang Yang
RAT combines repository-aware initialization, language-specific setup knowledge, and terminal-feedback agents to create runnable environments and verify repositories. On RATBench, RAT with DeepSeek-V3 reaches 63.2% Python ESSR, compared with Repo2Run’s 44.8%. Removing ImageRetriever lowers Python ESSR to 40.5%. This result concerns Python tests; for Java, Rust, JavaScript/TypeScript, and Go, success is based on build or installation commands, not test-suite pass rates.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.