Agent infrastructure dominated today. Backend branching, autonomous observability, and code review tools for AI-generated code all vying for attention alongside solid Mac utilities.
Microsoft Build dropped 7 in-house AI models, Alibaba shipped a multimodal agent, and Anthropic's Mythos program keeps finding thousands of zero-days across critical infrastructure.
Microsoft unveiled MAI-Thinking-1 and MAI-Code-1-Flash at Build, its first foundation models built entirely without OpenAI technology. MAI-Thinking-1 is a 35B-active-param MoE model scoring 97% on AIME 2025 and matching Claude Opus 4.6 on SWE-Bench Pro.
MAI-Code-1-Flash is rolling out to GitHub Copilot with a 16-point lead over Claude Haiku 4.5 on SWE-Bench Pro (51.2% vs 35.2%) and 60% fewer tokens. The full lineup includes MAI-Image-2.5 (#3 on Arena.ai), MAI-Transcribe-1.5 (#1 FLEURS benchmark), and MAI-Voice-2.