GitHub Copilot now lets you choose your favorite AI model

GitHub on Tuesday announced a major update to Copilot, its artificial intelligence (AI) coding assistance service. The announcement was made during the GitHub Universe 2024 event, which is being held in San Francisco. The update introduces multi-model support for Copilot, allowing developers to choose between different AI models from Anthropic, Google, and OpenAI. The company said the flexibility in choice will allow developers to freely use their preferred models for different projects. In addition, a new AI tool was also introduced, called GitHub Spark.

GitHub Copilot is getting an upgrade

Launched in 2021, GitHub Copilot was the first AI-powered platform to carry the Copilot branding. The AI ​​assistant was introduced just months after Microsoft invested in OpenAI and partnered with the AI ​​company. GitHub Copilot allows developers to use AI models to write code, help find bugs, perform debugging and security remediation, and more.

The event featured Microsoft’s encryption and file hosting platform introduced the GitHub Copilot will now offer developers a wider choice in the AI ​​models they want to use. Those using the AI ​​Assistant in Visual Studio Code and on the official website can now choose Claude 3.5 Sonnet from Anthropic, Gemini 1.5 Pro from Google and the GPT-4o, o1-preview and o1-mini models from OpenAI . While Claude 3.5 Sonnet is currently available, Gemini 1.5 Pro will be added in the coming weeks.

Developers can switch between models during a conversation with Copilot Chat to test and see which one suits them best. Users can also set a preferred AI model and start their project on it from the beginning.

GitHub Spark introduced

GitHub Spark is an AI-native tool that can be used by all developers, regardless of skill level, to generate micro-apps called a “spark.” These micro-apps are fully functional and can integrate AI capabilities and external data sources into larger apps, reducing dependency on cloud servers.

Generating a micro-app is also easy as developers can simply type a natural language prompt describing their requirements and previewing the app. Developers have the freedom to work directly on the app code to make the changes they want, or add follow-up prompts to let the AI ​​do the work. The tool supports both Anthropic and OpenAI models.

Once a spark is created, it can run automatically on desktops, tablets, or smartphones. Users can also share the spark with others – with custom access controls or by giving them full control to remix or build on the spark.