GitHub CoPilot provides Google and Entropic models

The CEO of GitHub has just announced that the programming assistant based on large language models, GitHub Copilot, will no longer be limited to OpenAI models, and users will be able to use models from other companies such as Google and Anthropic in the coming weeks. GitHub also unveiled the new Spark tool for building applications through prompts.

According to the CEO of GitHub, GitHub Copilot will soon become a multi-model tool and developers can choose different models from Google, Entropy and of course OpenAI. You can see this new feature in the image below:

According to the CEO of GitHub, in the next few weeks, Entropic’s Claude 3.5 Sonnet model will be released in Copilot, and Google’s Gemina 1.5 Pro model will be released a little later.

New models in CoPilot GitHub and the attractive feature of Spark

In addition to the Anthropic and Google models, GitHub will soon add support for a wider variety of OpenAI models, including GPT o1-preview and o1-mini. These models will be advanced from GPT-4 that Copilot used until now. Developers can switch between these models (even mid-chat) to find the model that suits their needs. Organizations can also choose which models team members use.

Of course, the new models will be added first in VS Code Copilot, but then in other sections such as Copilot Workspace, section for editing multiple files and code review, automatically fixing security issues and CLI, users can choose different models.

In addition to the addition of new models, GitHub also introduced Spark, a tool for developing natural language applications. Now, apart from coders, ordinary people can also build simple applications with natural language prompts; Of course, developers can still modify Spark production code more precisely. With the help of this feature, you can develop your desired program just by chatting, modifying and repeating your requests.

The new Spark feature also supports several Google, Entropic and OpenAI models. Spark is not yet available to most users, and those who want to try it out will have to join a waiting list.