The llama.cpp project has released version b9940, introducing an initialization for the `params.offline` parameter in `llama-bench`. This update is accompanied by a comprehensive set of pre-built binaries across multiple operating systems and hardware accelerators.

  • macOS Apple Silicon (arm64) and Intel (x64) builds are available, with KleidiAI support disabled.
  • Linux binaries cover Ubuntu x64 and arm64 CPU, Vulkan, ROCm 7.2, OpenVINO, and SYCL FP32/FP16.
  • Windows releases include CPU, OpenCL Adreno, CUDA 12.4/13.3, Vulkan, OpenVINO, SYCL, and HIP variants.
  • Android arm64 (CPU) and iOS XCFramework binaries are provided for mobile deployment.
  • openEuler builds for x86 and aarch64 architectures with ACL Graph support are included.

The release provides updated tooling for benchmarking offline parameters alongside extensive cross-platform compatibility for local model inference.