> For the complete documentation index, see [llms.txt](https://brindha.gitbook.io/mylearning/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://brindha.gitbook.io/mylearning/tools/magicoder.md).

# Magicoder

**What it is**

Magicoder is a series of fully open-source large language models for code that significantly closes the gap with top code models while having no more than 7B parameters. Magicoder models are trained on 75K synthetic instruction data using OSS-Instruct, a novel approach that uses open-source code snippets to generate diverse instruction data for code. 

**The OSS-Instruct method**

OSS-Instruct taps into the wealth of open-source code snippets online to generate diverse and realistic training data. The key idea is to overcome the inherent biases that can creep into synthetic data generated by AI models. By drawing on a vast repository of real-world code examples, Magicoder models can learn to write code that is more natural and applicable to real-world problems. 

**Performance**

Magicoder-S-DS-6.7B outperforms GPT-3.5-turbo-1106 and Gemini Ultra on HumanEval, scoring 76.8 against their 72.6 and 74.4 respectively. 

**Academic recognition**

Both Magicoder and MagicoderS substantially outperform state-of-the-art code models with similar or even larger sizes on a wide range of coding benchmarks. The research was accepted and presented at ICML 2024. 

**Developer**

Developed by researchers at the University of Illinois Urbana-Champaign (UIUC). Fully open-source including model code, weights, and training data.