Microsoft is training a new internal artificial intelligence (AI) model big enough to compete with those from Alphabet’s Google and OpenAI, The Information reported Monday.
The new model, known internally as MAI-1, is being overseen by recently hired Mustafa Suleyman, co-founder of Google DeepMind and former CEO of AI startup Inflection, according to the report, which cites two Microsoft employees with knowledge. of the initiative.
The exact goal of the model has not yet been determined and will depend on its performance. Microsoft could unveil the new model at its Build developer conference later this month, according to the report.
Microsoft declined to comment when contacted -.
MAI-1 will be “much larger” than previous smaller open source models that Microsoft had previously trained, meaning it will be more expensive, according to the report.
Microsoft last month launched a smaller AI model, called Phi-3-mini, that aims to attract a broader customer base with cost-effective options.
You may be interested in: Silicon Valley’s territorial war for the most experienced AI hacker house
Microsoft prepares a new AI model to compete with Google and OpenAI
The company has invested billions of dollars in OpenAI and deployed ChatGPT maker’s technology into its productivity software suite, allowing it to take the lead in the generative AI race.
According to the report, Microsoft has been setting aside large clusters of servers equipped with Nvidia graphics processing units along with large amounts of data to improve the model.
According to the report, MAI-1 will have about 500 billion parameters, while OpenAI’s GPT-4 will have one trillion and Phi-3 mini 3.8 billion.
In March, Microsoft named Suleyman head of its newly created consumer AI unit and hired several Inflection employees.
The new model does not come from Inflection, although it may be based on the startup’s training data, the report added.
With information from Reuters.
