San Francisco — GitHub, the widely-used cloud-based platform for developers, has introduced a new feature allowing users to interact with a variety of AI models, including GPT-4, Llama 3.1, and Mistral Large 2. This feature, known as “GitHub Models,” aims to make AI more accessible by integrating seamlessly with existing developer tools.

In a blog post, GitHub described the new “GitHub Models” feature as a way for developers to test different AI models and their parameters directly within their development environment. The feature is available for free and supports rapid prototyping and development of AI applications. GitHub has created a “glide path” for developers to transition from the playground to their development environments in Codespaces and VS Code.

GitHub’s interactive playground is designed not only for developers but also for students, startups, and hobbyists. It allows users to explore the latest AI models from providers like Meta, OpenAI, and Azure with just a single click. The platform assures users that it will not share outputs or prompts with AI model providers, nor will it use this data to train or enhance existing models.

The launch of “GitHub Models” positions GitHub in direct competition with platforms like HuggingFace, which also hosts AI models but lacks the same level of integration and advanced tools. The new feature is currently available in public beta, with developers needing to sign up for a waitlist to gain access.

Additionally, GitHub recently expanded its certification program, which now includes courses such as GitHub Foundations Certification and GitHub Actions Certification, further supporting the developer community.