The llama.cpp project has published build b10038, which includes an update to the release notes adding a link to the official website.

  • macOS and iOS binaries are available for Apple Silicon (arm64) and Intel (x64), along with an iOS XCFramework.
  • Linux builds support Ubuntu x64, arm64, and s390x architectures via CPU, Vulkan, ROCm 7.2, OpenVINO, and SYCL backends.
  • Android offers an arm64 CPU binary.
  • Windows provides x64 and arm64 binaries for CPU, OpenCL Adreno, CUDA 12/13, Vulkan, OpenVINO, SYCL, and HIP.
  • openEuler builds for x86 and aarch64 are listed, though some configurations like KleidiAI on macOS and certain openEuler options are marked as disabled.