The llama.cpp project has released version b10046, providing binaries for macOS, iOS, Linux, Android, Windows, and openEuler across various hardware backends including CPU, CUDA, Vulkan, ROCm, OpenVINO, SYCL, and HIP.

This release includes a fix to actually exercise the `test-recurrent-state-rollback` test. Additionally, KleidiAI support for macOS Apple Silicon has been disabled in this build.

The update is available for download from the official llama.cpp releases page.