
GitHub has launched GitHub Models, a feature that brings advanced AI models directly to its platform of over 100 million developers. This announcement comes just days after Microsoft's most recent earnings report, where CEO Satya Nadella shared that GitHub has reached an annual revenue run rate of $2 billion.
GitHub's growth is largely driven by its AI-powered developer tool, GitHub Copilot. In just over two years since its general availability, Copilot has been adopted by more than 77,000 organizations, including major players like BBVA, FedEx, and Infosys. This represents a 180% year-over-year increase in adoption.
The success of Copilot isn't just a win for GitHub's bottom line – it's reshaping how developers work. Copilot now accounts for over 40% of GitHub's revenue growth, highlighting the increasing demand for AI-assisted coding tools.
Building on this momentum, GitHub Models allows developers to experiment with and deploy top-tier AI models from Meta, Mistral, and Microsoft's Azure OpenAI service. The feature offers a built-in playground for testing various models, including Llama 3.1, GPT-4o, and Mistral Large 2.
What sets GitHub Models apart is its seamless integration into the development workflow. That means developers can transition from experimenting with prompts to deploying models in Codespaces and VS Code, and ultimately to production environments on Azure – all within the GitHub ecosystem.
This integration significantly lowers the barrier to entry for AI development, potentially accelerating the adoption of AI across a wide range of applications. GitHub is committed to privacy in this rollout, assuring that prompts and outputs won't be shared with model providers or used to improve the models.
Dohmke frames GitHub Models as part of a larger transformation in software development. The company's ambitious goal is to empower one billion developers – approximately 10% of the world's population – to build with AI.