Introducing StarCoder – The Best Free LLM for Developers

Since the launch of Chat-GPT in November last year, numerous companies have released a lot of Generative LLMs into the market. Natural Language Processing (NLP), a branch of AI that enables computers to understand written and spoken language like humans, makes these Large Language Models (LLMs) possible. In this blog, we will explore StarCoder, the best code LLM.

Let’s get started!

What is StarCoder?

Watch this!

StarCoder is a state-of-the-art large language for revolutionizing coding. It is a code LLM trained on licensed data from GitHub, encompassing code data from over 80 programming languages, Git commits, GitHub issues, and Jupyter notebooks. The usage of this training data was authorized to avoid legal issues akin to the ongoing lawsuits faced by OpenAI and Meta or GitHub’s lawsuit concerning the Copilot AI coding assistant.

Who developed this model?

ServiceNow and Hugging Face, the two companies under the BigCode Project, conceived StarCoder. The BigCode Project is an open scientific collaboration that actively works on responsibly developing and utilizing Code LLMs.

StarCoder vs Other LLMs

Evaluation of the StarCoder model using the HumanEval benchmark

StarCoder outperformed its larger counterparts when tested against the HumanEval benchmark. OpenAI released the HumanEval dataset, which includes 164 programming problems with a function signature, docstring, body, and several unit tests. They handwrote these problems to ensure their exclusion from the training set of code generation models. This exclusion became necessary to properly evaluate the model, as such models often undergo training using GitHub dumps. However, since OpenAI published the dataset on GitHub, there is a likelihood that it will be included in future dumps.

How does it work?

StarCoder Chat

You can chat using StarCoder from the Hugging Face website. They provide a very simple interface and anyone familiar with Chat-GPT can easily navigate it.

VSCode Extension

While being free to use, this extension can be a great coding assistant for your day-to-day coding tasks. From suggesting code blocks to code completion, this amazing tool is always at your service.


A lot is happening in the field of AI, and a revolution is imminent in the nearest future. One that’ll change the world as we know it now. How do you feel about this tool? Are you going to try it out?
Let us know your thoughts in the comments section.

Check out our blog on another Large Language model Med-PaLM 2.

Also, stay tuned for more updates.

Joy Mamah

Back to top