Monday, December 23, 2024

GitHub’s Copilot goes multi-model and adds support for Anthropic’s Claude and Google’s Gemini | TechCrunch

Must read

GitHub today announced that it will now allow developers to switch between a number of large language models when they use Copilot Chat, its code-centric ChatGPT-like service. Until now, Copilot Chat was powered by OpenAI’s GPT-4. Going forward, developers can choose between Anthropic’s Claude 3.5 Sonnet, Google’s Gemini 1.5 Pro, and OpenAI’s GPT-4o, o1-preview, and o1-mini. Anthropic will launch first, with Gemini following in a few weeks.

Some pundits may see this as yet another way for Microsoft to reduce its reliance on OpenAI, but GitHub CEO Thomas Dohmke framed it in terms of giving developers a choice. Not every model, after all, excels at every development-related task, and some models are simply better at working with certain languages than others.

“We truly believe that the era of a single model is over,” GitHub CEO Thomas Dohmke told me. “As you know, GitHub Copilot itself has always used multiple versions of the GPT-3 or -4 models for different tasks. Where low latency was required, we were using a faster, smaller model. Where we needed higher accuracy, we were using a bigger, more complex model. We, at GitHub, believe in developer choice and that developers — for reasons of company policy, benchmarks that they have seen, different programming languages and of course, personal preference, or because they’re using that model for other scenarios already — prefer one of the competing models, and so we’re officially partnering with both Anthropic and Google.”

The team had already done some of the prep work for this launch when it started offering developers the choice between GPT-4 and GPT-4 o1, which launched just over a month ago.

For now, this choice applies only to Copilot Chat and the newly launched Spark, but Dohmke noted that the company wants to bring this choice to all of its tools.

“Developers want a broad choice of models that are best-suited for development, including code generation, refactoring, and optimizing code,” said Google Cloud CEO Thomas Kurian. “Gemini models excel at this and are accessible on widely used developer platforms and environments – including now with GitHub Copilot – so millions of developers globally can benefit from trusted, enterprise-grade AI through Google Cloud.”

Latest article