llama.cpp Release b10897 Updates Windows on ARM CUDA Builds to CUDA 13.4.1 GA
Release b10897 of llama.cpp updates its build pipeline for Windows on ARM64 CUDA releases to use official CUDA 13.4.1 General Availability (GA) redistributables. This moves the Windows ARM64 build infrastructure away from Developer Preview archives. This update ensures stable, production-ready binaries for developers running LLMs locally on ARM-powered Windows devices with NVIDIA GPU acceleration. It aligns llama.cpp's CI pipeline with NVIDIA's official commercial CUDA support for Windows on ARM. The update specifically updates the binaries for `llama-b10897-bin-win-cuda-13.4-arm64.zip` and its associated CUDA DLL dependencies. The change was submitted via PR #28687 as a routine continuous integration maintenance update.
## BACKGROUND
llama.cpp is an open-source C/C++ inference engine for running large language models locally across various operating systems and hardware backends. NVIDIA recently expanded CUDA Toolkit support to include Windows on ARM, allowing PCs equipped with ARM CPUs and NVIDIA GPUs to leverage CUDA acceleration. Moving from Developer Preview packages to General Availability (GA) redistributables guarantees finalized, fully supported software binaries.