Microsoft is reportedly planning to integrate non-OpenAI models into its 365 Copilot platform, according to Reuters.
The move aims to reduce reliance on OpenAI, cut operational costs, and enhance speed for enterprise clients.
While Microsoft remains a key partner and backer of OpenAI, it is exploring internal and third-party AI models to support its AI assistant.
Microsoft stated that it uses various models from both OpenAI and its own development, depending on the product.
The company is also training its own AI model, Phi-4, and working to modify third-party models to further reduce costs.
Microsoft’s initial launch of 365 Copilot faced criticism, with customers raising concerns about its effectiveness, cost, and security. This push for diversified models could address these issues while strengthening Microsoft’s position in the competitive AI landscape.
Big Tech companies are heavily investing in AI startups to stay ahead. Amazon has partnered with Anthropic, investing $8 billion, while Google has acquired technology and talent from Character.AI.
Microsoft’s diversification strategy is a part of this broader race to develop advanced AI models.
Neither Microsoft nor OpenAI responded to requests for comment on the report.
