GitHub launches GitHub Models to integrate AI into workflows

GitHub has introduced a new feature called GitHub Models, aiming to provide developers with integrated access to a variety of AI models. This launch allows the platform's extensive user base, which exceeds 100 million developers, to utilise AI models directly within their workflows on GitHub.

The diversity of models included for initial access covers notable names in the AI industry, such as Meta, Mistral, OpenAI, Cohere, AI21 Labs, and Microsoft. This variety allows developers to experiment with different models and understand their capabilities without the need to leave GitHub. The playground feature offers a space for users to test different prompts and model parameters at no cost, providing an accessible way to explore both open and closed models.

Thomas Dohmke, CEO of GitHub, said, “Today, we democratise AI for the many. With GitHub Models, more than 100 million developers can now access and experiment with new AI models where their workflow is directly on GitHub. From OSS repositories to Copilot Extensions to GitHub Models, it is intentional that we bring as many partners into GitHub as possible. The world’s developers deserve to build with a wide array of world-class AI.”

GitHub Models will allow users to move beyond simple experimentation. Developers can integrate these AI models into workflows encompassing Codespaces, VS Code, and GitHub Actions, among other tools. GitHub aims to provide a cohesive platform where users can seamlessly learn, iterate, and deploy AI models within their existing development processes.

The new feature emphasises privacy and security; GitHub ensures that neither prompts nor outputs from GitHub Models will be shared with the model providers or used to retrain models. This aligns with GitHub and Microsoft’s commitment to maintaining user privacy and secure data handling practices.

Dohmke also outlined a broader vision for the future of development on GitHub, expressing a goal of enabling one billion developers, which would represent 10% of the world’s population, to build and advance technological progress. He stated, “In the years ahead, we will continue to democratise access to AI technologies to generate a groundswell of one billion developers. By doing so, we will enable 10% of the world’s population to build and advance breakthroughs that will accelerate human progress for us all.”

Developers will also benefit from a simplified pathway from experimentation to production. Using features like Codespaces and the GitHub CLI, developers can test code snippets in a variety of languages and frameworks. Upon readiness for production, Azure AI provides comprehensive tools like responsible AI practices, enterprise-grade security, data privacy, and substantial regional availability.

Anand Chowdhary, Co-founder of FirstQuadrant, highlighted the practical advantages of this integration: “As an AI startup founder and open source maintainer, GitHub Models enables my team to access and experiment with various LLMs in one place. This streamlines our development and lowers the entry barrier for building AI apps.”

The introduction of GitHub Models underscores GitHub’s strategy to become the central hub for AI development. By enhancing the capabilities of tools like GitHub Copilot, which already significantly augments coding efficiency, the company aims to make AI an integral part of the software development life cycle. According to GitHub, Copilot is currently responsible for writing nearly 50% of the code in files where it is enabled.

As of today, GitHub Models has entered a limited public beta phase. The service allows developers to sign up and start exploring the capabilities and potential applications of the AI models directly within their GitHub environment.