Microsoft is training a new in-house AI language model large enough to compete with those from Alphabet’s Google and ChatGPT owner OpenAI. Information reported Monday.
The new model, known internally as MAI-1, is being overseen by recently hired Mustafa Suleiman, co-founder of Google DeepMind and former CEO of AI startup Inflection, the report said. This is based on two Microsoft employees who are familiar with this initiative.
The exact purpose of the model has not yet been determined and will depend on how well it performs. Microsoft could preview the new model at the Build developer conference later this month, the report said.
Microsoft declined to comment when contacted by Reuters.
According to the report, MAI-1 will be “much larger” and more expensive than the previous, smaller open source model that Microsoft previously trained.
Last month, Microsoft launched a smaller artificial intelligence model called Phi-3-mini, hoping to reach a broader customer base with a cost-effective option.
The company invested billions in OpenAI and rolled out the ChatGPT maker’s technology across its productivity software suite, allowing it to take an early lead in the generative AI race.
According to the report, Microsoft has secured large clusters of servers powered by Nvidia’s graphics processing units and large amounts of data to improve its models.
According to the report, MAI-1 has approximately 500 billion parameters, while OpenAI’s GPT-4 has 1 trillion parameters, and Phi-3 mini is reported to measure 3.8 billion parameters. ing.
Microsoft appointed Mr. Suleiman to head its newly created consumer AI division in March and hired several Inflection employees.
The new model is not inherited from Inflection, but may be built on training data from the startup, the report added.
