vibehacker
News
arXiv ·

ProgramGym: 1,648 training envs for agents that rebuild programs from scratch

Xiamen University et al. open ProgramGym: 1,648 verifiable C/C++/Go/Rust environments built from GitHub repos for agents that rebuild programs from black-box behavior, plus a role-isolated Program-Agent harness. Fine-tuning Qwen3.8-27B on its trajectories raised ProgramBench from 34.87% to 42.84%; with Program-Agent the SFT model hit 56.25%.

More news

View all

Spotted something we missed? Start a thread.