← Back to live feed1 story
Microsoft introduced HydraFusion, a routing system that distributes AI tasks between on-device and cloud models within the Windows ecosystem to reduce operational costs. GitHub Copilot now routes routine coding work from Anthropic's Claude Haiku 4.5 to the locally running MAI-Code-1.1 Flash, a 137B parameter model with a 256K token context window, with on-device requests carrying no inference charges for users. The software depends on Nvidia RTX Spark hardware and Microsoft Execution Containers (MXC), now generally available on Windows 11, isolates agent activity to designated files and network locations for security.
Sign in to suggest edits
Key sources
- SOURCE@satyanadella“GitHub Copilot now hands off work to local models like MAI-Code-1.1 Flash, helping projects cost a lot less without sacrificing quality”x.com
- SOURCE@msdev“Coming soon, GitHub Copilot will determine when a task is best handled by on-device intelligence and when it should leverage cloud-scale models”x.com
- SUPPORT@mweinbach“$2599 for surface laptop ultra, Spark dev box $5999”x.com
- SUPPORT@abidlabs“The Windows launch event today has mentioned "local AI" 41 times (so far)”x.com
- SUPPORT@pavandavuluri“Microsoft Execution Containers (MXC) is now generally available on Windows 11, keeping agents contained within boundaries the operating system enforces”x.com
- SUPPORT@rohanpaul_ai“Copilot will hand routine coding from Anthropic's Claude Haiku 4.5 to Microsoft's MAI-Code-1.1-Flash (a 137B total and 6.8B active param model)”x.com
- SUPPORT@photoncap“up to 128GB unified memory and 1 petaflop FP4 performance”x.com
- SOURCEhuggingnewshuggingnews.com