Kimi-Dev-72B is a coding-specific LLM designed to autonomously resolve software issues and patch real repositories in Docker environments. The model is optimized via large-scale reinforcement learning, receiving rewards only when the entire test suite passes, ensuring solutions align with real-world development standards. It achieves state-of-the-art performance (60.4%) on SWE-bench Verified among open-source models and is available for download and deployment via Hugging Face.