
OpenAI's o1 model is now integrated into GitHub Copilot with 44% performance increase
The recently launched OpenAI's o1 model has been integrated into GitHub Copilot and GitHub Models, significantly enhancing developers' coding capabilities with a 44% performance boost over its predecessor, o1-preview.
The o1 model is available to Copilot Chat subscribers in Visual Studio Code and GitHub, where users can select the "o1 (Preview)" from the model picker. Subscribers to GitHub Copilot's Pro, Business, and Enterprise plans can access the model, though usage is capped at 10 messages every 12 hours. Business and Enterprise account administrators need to enable o1 access for team use. The model assists with various development tasks such as explaining, debugging, refactoring, testing, and code modernization, while integrating context from users' workspaces and GitHub repositories.
In GitHub Models, the o1 model aids in building AI applications through a playground, APIs, and model comparisons, allowing developers to experiment with o1 alongside models from Mistral, Cohere, Microsoft, and Meta.