Microsoft AI Updates: October 9, 2026

1. Windows Adds Local MAI Code 1.1 Flash, llama.cpp in Windows ML, and a Local-First Copilot

Microsoft. At its October 7 Windows event, Microsoft said MAI Code 1.1 Flash (137B total, 6.8B active parameters) is coming to devices at 3-bit precision with a 256K local context window, and that llama.cpp support has been added to Windows ML for running models across GPU, NPU and CPU. GitHub HydraFusion routing will extend to local models in the GitHub Copilot app, Copilot CLI and VS Code as an experimental preview later in October. The new Copilot, organized into Home, Code and Autopilot experiences, will gain local context, local actions across Windows and local model use on Copilot+ PCs starting in the coming months. Source

2. Surface Laptop Ultra and Surface RTX Spark Dev Box Open for Pre-Order

Microsoft. Microsoft opened pre-orders for the 15-inch Surface Laptop Ultra, built on the NVIDIA RTX Spark Superchip with up to 128GB of unified memory and up to 1 petaflop of FP4 AI performance for running models above 120B parameters locally, starting at $2,599 with availability on October 16. The Surface RTX Spark Dev Box, a compact desktop with the same compute and memory that ships with VS Code, GitHub Copilot, WSL, Python and Node preinstalled, is priced at $5,999 and ships in the U.S. in November. Microsoft also said upcoming Windows capabilities will use Microsoft Entra to distinguish agent activity from user activity and extend Agent 365 controls to local agents. Source