llama.cpp Release b11100 Fixes Tool-Call Parser Error for Muse Glimmer
llama.cpp release b11100 fixes a tool-call parsing error in chat mode when handling initial tool calls for the Muse Glimmer model family. The release also delivers updated pre-compiled binaries across macOS, Windows, Linux, Android, and Snapdragon platforms. This patch restores reliable function-calling execution for developers building autonomous AI agents with models like Muse Glimmer via llama.cpp. It ensures smooth compatibility and cross-platform binary availability across edge and high-performance hardware. The fix condensed trigger patterns and added automated tests in PR #29242 to ensure correct tool-call parser behavior in chat mode. Build binaries support various acceleration backends including CUDA 12/13, Vulkan, ROCm 10.0, OpenVINO, SYCL, and Snapdragon hardware.
## BACKGROUND
llama.cpp is a popular open-source C/C++ engine designed for efficient local inference of large language models across diverse hardware formats. Tool calling (or function calling) enables LLMs to structure their responses so external systems can execute code or query APIs. Muse Glimmer is a parameter-efficient model optimized for agentic and tool-use tasks.