HyperAI超神经

1. Tutorial Introduction

Kimi-Dev-72B is an open source large language model designed for software engineering tasks, released by the Chinese artificial intelligence company Dark Side of the Moon team on June 17, 2025. It achieved a performance of 60.4% in the SWE-bench Verified programming benchmark test, and won the championship with only 7.2 billion parameters, surpassing the recently released new version of DeepSeek-R1 with 67.1 billion parameters, becoming the SOTA among the current open source models.

Key features:

Code Repair (BugFixer): Automatically locate errors in the code and generate repair patches to solve vulnerabilities and defects in software development.
Test Code Generation (TestWriter): Write unit tests for existing code to ensure code quality and functional correctness.
Automated development process: Based on reinforcement learning and self-game mechanism, it coordinates repair and testing capabilities to improve development efficiency.
Integration with development tools: In the future, it will be deeply integrated with IDE, version control systems, and CI/CD pipelines, and seamlessly integrated into the development workflow.

Technical principle:

Division of roles (BugFixer and TestWriter): The model plays two roles, responsible for fixing code and writing tests respectively, and both share the framework of file location and code editing.
Mid-training: Use approximately 150 billion high-quality data points for training to enhance the model’s understanding of actual development tasks.
Reinforcement Learning: Run the code in a Docker environment and give rewards based on the test results to improve the model's code editing capabilities.
Test-time Self-Play: During the test phase, the model generates multiple patches and test candidates, and coordinates repair and testing capabilities based on a self-play mechanism to improve overall performance.

Application scenarios:

Code repair: Quickly detect and repair errors or vulnerabilities in the code, reducing the time for manual investigation and repair.
Test code generation: Generate high-quality unit test code for existing code to improve test coverage.
Development process automation: Integrates with mainstream IDEs to provide real-time code repair suggestions and test code generation functions.
Programming education: Help beginners quickly understand and master programming skills by generating sample code and test code to assist learning.
Open source project maintenance: Help maintainers of open source projects quickly fix vulnerabilities and optimize codes to improve project quality and stability.

The computing resources of this tutorial use a single RTX A6000 card. The model deployed in this tutorial is Kimi-Dev-72B-GGUF/Kimi-Dev-72B-IQ4_NL.gguf.

2. Effect display

3. Operation steps

1. Start the container

2. After entering the webpage, you can start a conversation with the model

If "Model" is not displayed, it means the model is being initialized. Since the model is large, please wait about 2-3 minutes and refresh the page.

4. Discussion

🖌️ If you see a high-quality project, please leave a message in the background to recommend it! In addition, we have also established a tutorial exchange group. Welcome friends to scan the QR code and remark [SD Tutorial] to join the group to discuss various technical issues and share application effects↓

Citation Information

The citation information for this project is as follows:

@misc{kimi_dev_72b_2025,
  title        = {Introducing Kimi-Dev-72B: A Strong and Open Coding LLM for Issue Resolution},
  author       = {{Kimi-Dev Team}},
  year         = {2025},
  month        = {June},
  url          = {\url{https://www.moonshot.cn/Kimi-Dev}}
}