According to a Bloomberg report today, Microsoft has independently developed artificial intelligence (AI) models and believes they can compete with industry leaders, including OpenAI, a company it partners with.
Sources familiar with the matter revealed that Microsoft has recently tested a series of self-developed AI models, showing performance comparable to state-of-the-art AI solutions from OpenAI and Anthropic.
The company is evaluating these models, called MAI, across different AI tasks and is also working on more advanced inference models designed to handle complex problem-solving with near-human reasoning capabilities.
A Microsoft spokesperson stated: “We leverage multiple AI models, including those from our deepening collaboration with OpenAI, as well as our own and open-source AI solutions.”
During a Morgan Stanley conference, Microsoft CFO Amy Hood emphasized:
“Our successes are mutually reinforcing. In AI, we are planning for the next 10 to 20 years, which is crucial for both sides.”
Currently, Microsoft has launched self-developed small AI models, Phi, and sells AI solutions from various companies. Reports indicate that Microsoft has tested AI models from Anthropic, DeepSeek, Meta, and xAI to assess their potential within its ecosystem.
“We remain very proud of OpenAI’s leading models, but we are also developing and integrating other AI solutions to provide users with more choices,” Hood added.
