llama.cpp Release b11531 Refactors Chat API and Updates Multi-Platform Binaries
llama.cpp released automated build release b11531, featuring a refactor of the internal chat API via Pull Request #30210. The release includes updated pre-built executable binaries across desktop, mobile, and specialized computing platforms. Continuous refactoring of core components like the chat API helps maintain clean code abstractions as llama.cpp evolves. The automated build pipeline ensures users across macOS, Linux, Windows, and Android can deploy the latest LLM inference improvements on various hardware acceleration backends. This release provides binaries supporting CPU and GPU backends including CUDA 12/13, Vulkan, ROCm 10.0, SYCL, OpenVINO, and Qualcomm Snapdragon (Adreno GPU / Hexagon NPU). Notably, the Apple Silicon build with ARM KleidiAI integration remains disabled.
## BACKGROUND
llama.cpp is a widely used open-source C/C++ library for running Large Language Models locally with minimal setup and high hardware efficiency. ARM KleidiAI is an open-source library developed by ARM to accelerate low-level AI inference operations on ARM-based processors.