--- title: README emoji: 🦊 colorFrom: yellow colorTo: red sdk: static pinned: false --- # ProgramBench: Can Language Models Rebuild Programs From Scratch? Given only a compiled binary and its documentation, AI agents must architect and implement a complete codebase that reproduces the original program's behavior. ProgramBench evaluates this capability across 200 real-world open-source projects spanning Rust, Go, C, C++, Haskell, and Java. ## Links - [Website & Leaderboard](https://programbench.com) - [Paper](https://programbench.com/static/paper.pdf) - [GitHub](https://github.com/facebookresearch/ProgramBench) - [Documentation](https://programbench.com/more)