Today, GitHub announced GitHub Models, a new service that allows developers to find and experiment with AI models for free. It brings the power of leading large and small language models directly to GitHub"s 100+ million users.
GitHub Models will provide access to leading models, including OpenAI"s GPT-4o and GPT-4o mini, Microsoft"s Phi 3, Meta"s Llama 3.1, and Mistral Large 2. In the coming months, GitHub will be adding more language and vision models as they approach general availability.
Developers can access these models via a built-in playground where they can test different prompts and model parameters. From the playground, developers can move their code to Codespaces and VS Code for further development. Once ready for production with Azure AI, they can simply replace their GitHub personal access token with an Azure subscription and credentials.
GitHub explicitly mentioned that no prompts or outputs from GitHub Models will be shared with model providers and will not be used to train or improve any AI models.
Thomas Dohmke, GitHub CEO, wrote the following regarding the GitHub Models launch:
GitHub Copilot is foundationally changing the speed of software production, already writing nearly 50% of code in files where it’s enabled. With GitHub Copilot Workspace, we envision a world where millions of novice, hobbyist, and professional developers alike can code with entirely human language.
And now with GitHub Models, more than 100 million developers can access and experiment with new AI models where they already manage their source code, issues, pull requests, workflows, and repositories – directly on GitHub.
Microsoft CEO Satya Nadella tweeted the following regarding the GitHub launch:
Introducing GitHub Models, a new way for the more than 100 million developers who call GitHub home to build with industry-leading AI models directly from their workflow. https://t.co/zT6zPeiKX8
— Satya Nadella (@satyanadella) August 1, 2024
You can sign up for the limited public beta of GitHub Models here.