GitHub Copilot will move from exclusively using OpenAI’s ChatGPT models to a multi-modal approach. The company announced this new adoptive measure at the annual GitHub Universe Conference on Tuesday, October 29.
GitHub CEO Thomas Dohmke in a blog post revealed that they are bringing developer choice to GitHub Copilot with Anthropic’s Claude 3.5 Sonnet, Google’s Gemini 1.5 Pro, and OpenAI’s o1-preview and o1-mini.
Since the launch of GitHub Copilot using OpenAI Codex, the AI-powered coding assistant has gone through significant updates on the base model versions from GPT 3.5-turbo to GPT 4o and 4o-mini models. The CEO stated that the next phase of AI code generation will not only be defined by multi-model functionality but also by multi-model choice.
The new models are available starting with Copilot Chat, and then Anthropic’s Claude 3.5 Sonnet and Google Gemini 1.5. Copilot Chat is available now with OpenAI o1-preview and o1-mini, Sonnet will be available in the coming week and Gemini 1.5 over the upcoming weeks.
The CEO also assured bringing various functions and multi-model choices on the GitHub Copilot’s surface ranging from Copilot Workspace, multi-file editing, code review, security autofix, and the CLI.
Many other changes are made on the GitHub Copilot including extensions, the ability to create various files from a chat with Visual
Other than the major announcement of broadening the models beyond OpenAI models, GitHub announced the launch of an AI tool called Spark for building web apps. They are also launching updates to GitHub Copilot in Visual Studio Code, Copilot for Xcode, and more.
GitHub users in the coming week can select Claude 3.5 with Gemini 1.5 using their web or VS Code. Developers can choose from different models and work on them at the same time while remaining in the conversation with Copilot Chat. This will let the developers give plenty of choices to choose from that suit their interests and tasks.
The CEO Thomas Dohmke believes that there is no one model to rule every scenario, and developers expect the agency to build with the models that work best for them.
Microsoft aimed to increase its use and bring developer freedom, openness, and innovation when it bought GitHub in 2018 and the revenue reached $1 billion in 2022. GitHub announced that Copilot reached a milestone of 1 million paid subscribers last week and the latest multi-model approach will open up more possibilities.
The new Claude 3.5 Sonnet can be used for various coding tasks, like base designing, bug fixes, maintenance, and optimizations. It can do complex and multi-step coding, app updates, code refactoring, and feature development.
The Gogle’s model Gemini 1.5 Pro is built as a multi-model with high coding capacities. It features a two million token context window, that can process code, images, audio, video, and text simultaneously.
This is the longest of any large-scale foundation model and developers can process large lines of codes, more than 100,000 lines. Gemini 1.5 Pro also has an impressive response time and works well for common developer uses like code generation, analysis, and optimization.
The AI models from OpenAI used for GitHub started with GPT 3 are extending to o1-preview and o1-mini which are more advanced than GPT 4o. The new models are updated to increase the understanding level of code constraints and edge cases with greater reasoning capabilities. This will result in an efficient and quality coding.
With the development of GitHub Spark, GitHub is broadening its field as well as users from a million to a billion in the coming years. It was introduced at the Universe Conference and is an AI-native tool to build applications entirely in natural language, said Dohmke in the blog post. He explained, “Sparks are fully functional micro apps that can integrate AI features and external data sources without requiring any management of cloud resources.”
Other Articles: Google’s Gemini 2.0 AI Model May Launch in December
