GitHub debuts tool for AI developers to experiment with Gen AI

GitHub has introduced a new feature known as GitHub Models that makes generative AI models more accessible to developers, enabling them to try out and build applications. Like all the other GitHub’s new features, GitHub Models is available in limited public beta.

According to Forbes, the new capabilities allow developers to access and use different models within the GitHub work interface. With the new GitHub Models, developers can compare and sample different models without leaving that current environment.

The new feature provides a playground for developers

The latest product enables interaction with existing models like Meta’s Llama 3.1, OpenAI’s ChatGPT-4o and GPT-4o mini, Cohere’s Command, and Mistral Large 2 by Mistral AI.

According to Forbes, this integration with leading AI models is critical for AI developers as it gives them an opportunity to experiment with various models directly within GitHub Codespaces, which will accelerate the process of developing AI applications from prototype to production.

“We are enabling the rise of the AI engineer with GitHub Models–bringing the power of industry-leading large and small language models to our more than 100 million users directly on GitHub,” said GitHub in a blog post.

“And if you like what you’re seeing on the playground, we’ve created a glide path to bring the models to your developer environment in Codespaces and VS Code.”

GitHub.

The new feature closely resembles the capabilities of the model playground that is offered by model and cloud providers including Microsoft Azure. ChatGPT maker OpenAI was reportedly one of the first firms to launch a playground that allowed users to test different models and generate code based on the configuration.

Now, Github is enabling similar features by integrating its command line tools – Codespaces and Visual Studio Code.

GitHub Models falls in line with Microsoft’s plans

The initiative aspires to cut on the AI friction for developers by providing a centralized catalog of AI models they can access and test within the GitHub platform using their existing GitHub identity.

Mario Rodriguez, senior vice-president of product at GitHub, told VentureBeat that it was previously impossible for GitHub’s users to easily access and explore an array of gen AI models using just a GitHub identity.

However, developers that use GitHub can now access an array of capabilities making it easier to develop code.

“We just wanted to make it extremely simple, you know, AI is not a fad, it’s here to stay. So, we just have to get that friction to be zero, if we want to continue to have that market grow.”

Rodriguez.

This also makes GitHub an alternative to platforms such as Hugging Face, which, according to Forbes, lacks the deep integration with development tools that GitHub provides.

Now, GitHub is bridging the gap, allowing developers to experiment with AI models while also quickly exporting and integrating their code into existing workflows.

The initiative is also in line with Microsoft’s strategy to enhance the accessibility and usability of AI as it allows developers to transition to Azure to scale their solutions. According to Forbes, this integration pushes Microsoft’s goal of providing a comprehensive, developer-friendly path from experimentation deployment.