Microsoft announced new AI capabilities for Windows, including MAI-Code-1.1 Flash, a 137B parameter coding model optimized for local PC execution. GitHub Copilot now leverages this model to reduce costs and improve security by keeping sensitive work on the device.
Key facts
| Fact | Detail | The source says |
|---|---|---|
| MAI-Code-1.1 Flash Parameters | 137B | “MAI-Code-1.1 Flash: 137B parameter coding model w/ 256K context window, which is now optimized to run on your PC!” |
| GitHub Copilot Integration | Local model execution | “GitHub Copilot now hands off work to local models like MAI-Code-1.1 Flash, helping projects cost a lot less without sacrificing quality.” |
| Hybrid Intelligence | Secure local execution | “With Hybrid Intelligence, Copilot can now take action directly on the PC and keep sensitive work on your device.” |
| Device Power | NVIDIA RTX Spark | “Surface Laptop Ultra, powered by NVIDIA RTX Spark.” |
What happened
Microsoft today announced a new chapter for Windows, focusing on bringing AI capabilities directly to PCs. Key highlights include the introduction of MAI-Code-1.1 Flash, a 137B parameter coding model optimized to run locally on PCs. GitHub Copilot now utilizes this model to reduce costs and improve security by keeping sensitive work on the device. With Hybrid Intelligence, Copilot can take action directly on the PC, ensuring secure boundaries for agent execution. Microsoft also emphasizes the importance of security, integrating Windows and Agent 365 to create a secure environment for AI agents. The new generation of devices, like the Surface Laptop Ultra, will be powered by NVIDIA RTX Spark, enabling these advanced AI capabilities.
In the Fathom Layer index
- GitHub Copilot — Developer Tools
What to weigh
- MAI-Code-1.1 Flash is optimized for local execution.
- Hybrid Intelligence ensures secure local execution.
- Local execution reduces cloud token spend.
FAQ
What is MAI-Code-1.1 Flash?
MAI-Code-1.1 Flash is a 137B parameter coding model optimized to run on your PC.
How does GitHub Copilot benefit from local models?
GitHub Copilot now hands off work to local models like MAI-Code-1.1 Flash, helping projects cost a lot less without sacrificing quality.
What is Hybrid Intelligence?
With Hybrid Intelligence, Copilot can now take action directly on the PC and keep sensitive work on your device.
What powers the new generation of devices?
The new generation of devices, like the Surface Laptop Ultra, will be powered by NVIDIA RTX Spark.
Source: Microsoft