News
'GitHub Models' AI Playground Debuts with Llama 3.1, GPT-4o and More
GitHub today previewed GitHub Models to help its 100-million-plus developers try out the latest/greatest AI models from a range of providers.
Developers can use it to build their understanding of AI model capabilities by experimenting with model settings and sending prompts through a chat interface, or they can directly interact with models through an SDK. GitHub users can sign up now for a limited public beta.
"From Llama 3.1, to GPT-4o and GPT-4o mini, to Phi 3 or Mistral Large 2, you can access each model via a built-in playground that lets you test different prompts and model parameters, for free, right in GitHub," the Microsoft-owned company's Thomas Dohmke announced today, Aug. 1. "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. And once you are ready to go to production, Azure AI offers built-in responsible AI, enterprise-grade security & data privacy, and global availability, with provisioned throughput and availability in over 25 Azure regions for some models. It's never been easier to develop and run your AI application."
[Click on image for larger view.] GitHub Models Site (source: GitHub).
He described the initial offering as "just the first wave" and promised to add more language, vision, and other models on the months ahead as general availability approaches.
Dohmke explained how the new offering helps developers test and compare different models and leverage the customized abilities of each.
"Every piece of software is unique," he said. "And likewise, every model is unique in its capabilities, performance, and cost. Mistral offers low latency, while GPT-4o is excellent at building multimodal applications that might demand audio, vision, and text in real time. Some advanced scenarios might require the integration of different modes, such as an embeddings model for Retrieval Augmented Generation (RAG).
He described the initial offering as "just the first wave" and promised more language, vision and other models to be added in the months ahead as general availability approaches.
[Click on image for larger view.] GitHub Models Screenshot(source: GitHub).
GitHub also provided guidance on prototyping with AI models and Responsible use of GitHub Models.
The latter explains GitHub Models is designed to allow for learning, experimentation and proof-of-concept activities and is subject to various limits, such as requests per minute, requests per day, tokens per request, and concurrent requests. GitHub said it's not designed for production use cases. Safeguards include a number of content filters that can't be turned off.
Also, to align with GitHub and Microsoft's commitment to privacy and security, no prompts or outputs in GitHub Models will be shared with model providers, nor used to train or improve the models.
The new offering joins a host of "playgrounds" provided by GitHub's owner, Microsoft, as indicated by articles including:
Developers can go here to sign up for the new offering, though they are warned, "Admission to the limited public beta for GitHub Models will be limited. You will receive an email if you are granted access."
About the Author
David Ramel is an editor and writer at Converge 360.