On December 24th, sources familiar with the matter revealed that Microsoft has been working to increase internal and third-party artificial intelligence models to power its flagship AI product, Microsoft 365 Copilot, in order to break away from the current underlying technology from OpenAI and reduce costs.
Insiders say that Microsoft is also seeking to reduce 365 Copilot's reliance on OpenAI due to concerns about the cost and speed for enterprise users. A Microsoft spokesperson stated that OpenAI will continue to be the company's partner in cutting-edge modeling. The original agreement between the two companies allows software giants to customize OpenAI's models.
Insiders added that in addition to training its own small models (including the latest Phi-4), Microsoft is also working hard to customize other open heavyweight models to make 365 Copilot faster and more efficient. (Reuters)