Microsoft ditches OpenAI dependency with first in-house reasoning model
microsoft revealed mai-thinking-1, its first in-house reasoning model with 35b parameters, signaling a shift away from exclusive reliance on openai as it expands its autonomous agent ecosystem.
microsoft revealed mai-thinking-1, its first in-house reasoning model featuring 35 billion active parameters and a 128k context window optimized for complex logic and code. it simultaneously launched scout, an autonomous 'always-on' assistant built on the open openclaw platform to handle background tasks like calendar management and reporting. these announcements, alongside new developer hardware and an os for cross-device agents, signal that microsoft is moving to secure its own ai stack rather than relying solely on openai.