Microsoft is Building 'MAI-1' to Compete With OpenAI and Google AI Models: Report MAI-1 will be larger than previous open-source models developed by the Satya Nadella-led tech giant
You're reading Entrepreneur India, an international franchise of Entrepreneur Media.
Microsoft, an investor in OpenAI and a competitor to Google, is reportedly working on getting its Large Language Model to the market to rival GPT-4 and Gemini.
The model is being led by Google Deepmind founder and CEO of AI startup Inflection, Mustafa Suleyman. In March, Microsoft hired a considerable number of the startup's employees and bought its Intellectual Property rights for USD 650M.
As reported by The Information, the model is not carried forward from Inflection, but the model may have been developed on the training data and other tech from the startup. Two Microsoft employees familiar with the development said this model is different from the previous ones developed by the startup.
Microsoft could preview the new model as soon as its Build developer conference later this month.
MAI-1 will be larger than previous open-source models developed by the Satya Nadella-led tech giant.
Reportedly, MAI-1 will have 500 billion parameters, as compared to that of OpenAI's GPT-4's 1 trillion parameters. While smaller open-source models released by firms like Meta Platforms and Mistral have 70 billion parameters.