llama.cpp b9960 removes loading.html, streamlines UI
New release cuts startup bloat and updates interface for local LLM runners.
The latest release of llama.cpp, version b9960, brings a targeted UI refinement: the removal of loading.html and associated UI changes (PR #25500). This move streamlines the initial page load for users running the project's built-in web server, reducing unnecessary bloat and simplifying the codebase. The change was verified with a GPG signature, ensuring integrity.
With b9960, llama.cpp continues its tradition of broad platform support. Pre-built binaries are available for macOS (Apple Silicon including KleidiAI-accelerated builds, Intel), iOS (XCFramework), Linux (x64/arm64 with CPU, Vulkan, ROCm 7.2, OpenVINO, SYCL), Android (arm64), and Windows (x64/arm64 with CPU, CUDA 12/13, Vulkan, OpenCL Adreno, OpenVINO, SYCL, HIP). This release is ideal for developers deploying local LLMs with minimal overhead.
- Removed loading.html and applied UI changes, reducing startup bloat.
- Supports 27 build targets across macOS, Windows, Linux, Android, and iOS.
- Acceleration options include CUDA 12/13, Vulkan, ROCm, OpenVINO, SYCL, and KleidiAI on Apple Silicon.
Why It Matters
A leaner web UI means faster local LLM deployments and a cleaner developer experience for open-source AI users.