$Microsoft (MSFT.US)$ is reportedly gearing up to compete with tech giants $Alphabet-A (GOOGL.US)$ Google and OpenAI by developing a new in-house AI language model.
Microsoft, under the supervision of Mustafa Suleyman, former Google DeepMind co-founder, is training a new AI language model named MAI-1, according to a report from the Information.
This initiative seeks to rival the capabilities of Google and OpenAI. The purpose of MAI-1 is yet to be fully determined and is contingent upon its performance.
The model's unveiling may occur during Microsoft's upcoming Build developer conference, the report added.
MAI-1 is projected to surpass Microsoft's previous smaller, open-source models in size and capability, potentially signaling higher costs.
The company recently introduced Phi-3-mini, a smaller AI model, aiming to broaden its client base with cost-effective alternatives.
Phi-3-mini, measuring 3.8 billion parameters, is available in the Microsoft Azure AI Model Catalog and on Hugging Face, a platform for machine learning models, as well as Ollama, a lightweight framework for running models on a local machine. It will also be available as an $NVIDIA (NVDA.US)$NIM microservice with a standard API interface that can be deployed anywhere.
Microsoft's substantial investment in OpenAI technology and deployment of ChatGPT across its suite of productivity software has positioned it as a frontrunner in generative AI.
The report noted that MAI-1 is anticipated to boast approximately 500 billion parameters, in contrast to OpenAI's GPT-4 with one trillion parameters and Phi-3-mini with 3.8 billion parameters.
Microsoft stock has gained more than 33% in the past 12 months.
Disclaimer: This content was partially produced with the help of AI tools and was reviewed and published by Benzinga editors.