Advancing the Quality-Efficiency Frontier for Coding AI on Windows PCs - Microsoft Foundry on Windows Blog Coding AI on Windows: The Quality–Efficiency Frontier

Microsoft has introduced MAI-Code-1.1-Flash, a 137.9-billion-parameter mixture-of-experts coding model optimized to run locally on Windows PCs with NVIDIA RTX Spark hardware. By combining mixed-precision quantization, speculative decoding via a DFlash2 draft model, and bounded sliding-window attention for long contexts, the system achieves significant improvements in memory footprint, decoding speed, and long-context efficiency. The model is made available through Microsoft Foundry and as a local model in GitHub Copilot. Open article →