B, an open-weight multimodal vision AI model designed to deliver strong math, science, document and UI reasoning with far less training data and compute than much larger systems.

Microsoft's Phi-4-reasoning-vision-15B uses careful data curation and selective reasoning to compete with models trained on five times more data, reshaping the small AI playbook.

Microsoft's AI image generator offers impressive realism and text rendering, but strict content limits and 1:1-only output hold it back.

Microsoft has released two new Canary builds, 28020.1743 under KB5081339, and 29553.1000, for Windows 11 Insiders with ...

AI agents will become part of the enterprise application layer. Microsoft’s new Agent 365 platform signals a shift from AI as a chat assistant to AI as an operational actor capable of calling services ...

YouTube Live has added the Nimbra Edge and Nimbra 400 series to its list of verified encoders. Inclusion on the list confirms that YouTube has evaluated the encoding solutions for live encoder ...

DirectStorage 1.4 brings along key upgrades to the API, including support for Zstandard compression as well as CreatorID for ...

Apple's iOS 26.4 beta adds AI-generated Apple Music playlists, early testing for encrypted RCS messaging and a new setting to tone down brighter visual effects. With more than a decade of experience, ...

TL;DR: If you're looking for a powerful productivity suite without recurring subscription fees, Microsoft Office 2021 is available for just $39 for a limited time, an incredible deal. This offer ...

For better or worse, Microsoft Teams is one of the most important communication apps in the professional world. It’s used by millions for chat messages, video conferences, and sending files. According ...

I’ve been waiting for Microsoft’s Surface Pro 5 for what seems like years. Now I know that my wait will never end. There will be no Surface Pro 5. There is a Surface Pro. On Tuesday in China, ...

IBM has released Granite 4.0 1B Speech, a compact 1-billion-parameter multilingual speech model that ranks first on OpenASR with a 5.52 Word Error Rate.