By Adam Clark

Microsoft and Google-parent Alphabet are slugging it out in the field of artificial intelligence. The latest arena is adding more reasoning capabilities to their AI models as they look to justify more expensive subscriptions to productivity suites.

Microsoft said it would launch two "first-of-their-kind reasoning agents" which it calls Researcher and Analyst, for customers of its Microsoft 365 Copilot software, late on Tuesday. AI agents are programs that have the ability to take simple directions and complete multistep tasks.

"They analyze vast amounts of information with secure, compliant access to your work data -- your emails, meetings, files, chats, and more -- and the web to deliver highly-skilled expertise on demand," wrote Jared Spataro, Microsoft's chief marketing officer for AI at Work, in a blog post.

Access to the agents will begin in April. Microsoft noted that is part of a new program to give customers early access to new Copilot features while they are still in development. Copilot is available for business customers as a $30 a month add-on to Microsoft 365 subscriptions

The move underlines the fierce competition between AI providers hoping to convince users to pay up for access to better models and justify their investment in AI infrastructure. Reasoning models use as much as 100 times more computing capacity than previous AI models, according to Nvidia CEO Jensen Huang.

On the same day as Microsoft's release, Google said it had released its own updated AI model called Gemini 2.5, also integrating more reasoning abilities. It said that on a variety of benchmarks it beat AI models from Microsoft-backed OpenAI, Elon Musk's xAI and China's DeepSeek.

The new model is available in Google AI Studio for developers and for Gemini Advanced users -- available via a paid subscription plan. The company said it would be added to its Vertex AI product for cloud-computing customers soon and it will release information on pricing in the coming weeks.

"Going forward, we're building these thinking capabilities directly into all of our models, so they can handle more complex problems and support even more capable, context-aware agents," wrote Koray Kavukcuoglu, chief technology officer at Google DeepMind, in a blog post.

Write to Adam Clark at adam.clark@barrons.com

This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.

(END) Dow Jones Newswires

March 26, 2025 07:55 ET (11:55 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.