b9260

Release b9260 of llama.cpp includes OpenCL backend refactoring that improves initialization, GPU identification, and performance by caching global memory size in device context. llama.cpp enables LLM inference with minimal setup and high performance across diverse hardware platforms locally and in the cloud. Pre-built binaries and source code are available for multiple platforms including macOS, Linux, and Windows.

Source: llama.cpp Releases | 2026-05-21