Today, GitHub announced a new service, GitHubModels, that allows developers to find and try out artificial intelligence models for free. It brings the power of leading large and small language models directly to GitHub’s more than 100 million users.

GitHub Models will provide access to leading models, including OpenAI’s GPT-4o and GPT-4omini, Microsoft’s Phi3, Meta’s Llama3.1 and MistralLarge2. In the coming months, GitHub will continue to add more language and visual models.

Developers can access these models through a built-in playground, where they can test different cues and model parameters. Developers can transfer their code from the gaming floor into Code Space and VSCode for further development. Once they are ready to use Azure AI for production, they simply replace their GitHub personal access token with their Azure subscription and credentials.

GitHub clearly mentions that tips or output from GitHub models will not be shared with model providers, nor will they be used to train or improve any artificial intelligence models.

GitHub CEO Thomas Dohmke wrote the following about the release of GitHubModels:

GitHub Copilot is fundamentally changing the speed of software production, and it already accounts for nearly 50% of the code in files that enable the feature. With GitHub Copilot Workspace, we hope millions of novice, hobbyist, and professional developers can write code in a completely human language.

Now, with GitHub Models, more than 100 million developers can manage their source code, issues, pull requests, workflows, and repositories directly on GitHub, and access and experiment with new artificial intelligence models.

Microsoft CEO Satya Nadella commented on the GitHub launch on Twitter:

GitHub Model is a new way for the more than 100 million developers who call GitHub home to build directly in their workflows using industry-leading artificial intelligence models.

You can sign up for the limited public beta of GitHubModels here.