The llama.cpp project has added conversion support for Granite Speech Plus, extending its capability to handle multi-layer concatenation. This update is available across multiple platforms including macOS, Linux, Android, Windows, and openEuler, with support for various hardware accelerators and AI models.