Download
CodeGeeX2 is the second-generation multilingual code generation model from ZhipuAI, built upon the ChatGLM2-6B architecture and trained on 600B code tokens. Compared to the first generation, it delivers a significant boost in programming ability across multiple languages, outperforming even larger models like StarCoder-15B in some benchmarks despite having only 6B parameters. The model excels at code generation, translation, summarization, debugging, and comment generation, and it supports over 100 programming languages. With improved inference efficiency, quantization options, and multi-query/flash attention, CodeGeeX2 achieves faster generation speeds and lightweight deployment, requiring as little as 6GB GPU memory at INT4 precision. Its backend powers the CodeGeeX IDE plugins for VS Code, JetBrains, and other editors, offering developers interactive AI assistance with features like infilling and cross-file completion.
Features
- Trained on 600B code tokens with significant multilingual coding improvements
- Surpasses larger models (e.g., StarCoder-15B) on benchmarks like HumanEval-X
- Efficient inference with Flash Attention and Multi-Query Attention
- Supports 100+ programming languages with infilling and cross-file completion
- Lightweight deployment with INT4 quantization needing ~6GB GPU memory
- Integrates into IDEs (VS Code, JetBrains, PyCharm, WebStorm, etc.) via CodeGeeX plugin
Project Samples
