Models & Labs
llama.cpp b9128 Release Optimizes Hexagon and macOS Support
llama.cpp Releaseshigh confidence
Why it matters
- →Optimizes performance for Hexagon by eliminating scalar VTCM loads.
- →Enhances macOS support, including Apple Silicon with KleidiAI.
- →Expands compatibility across diverse hardware configurations.
The b9128 release of llama.cpp introduces optimizations for Hexagon, focusing on eliminating scalar VTCM loads through HVX splat helpers. This update also enhances support for macOS, including Apple Silicon with KleidiAI enabled, and extends compatibility across multiple platforms such as Windows and Linux. Key improvements include optimized per-group scale handling and slope load from VTCM. These enhancements aim to boost performance and efficiency, making llama.cpp more adaptable for developers working with various hardware setups.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1Llama.cpp b9082 Release Enhances Hexagon Backend
llama.cpp Releases · May 9, 2026 · Same story
What happened next
- 2Llama.cpp b9946 Release Enhances Hexagon Optimizations
llama.cpp Releases · July 11, 2026 · Same story
More from llama.cpp Releases
Coding Toolscoding
llama.cpp b12022 adds ROCm 10 and Snapdragon support
This release quietly cements llama.cpp as the universal inference runtime by finally bringing ROCm 10.0 to Linux and Windows alongside CUDA 13.4, effectively closing the hardware gap for AMD users who were previously left behind. The inclusion of native Snapdragon support on Linux marks a significant step toward running local LLMs on ARM-based mobile chips without emulation overhead. While KleidiAI on Apple Silicon is disabled in this build, the broader expansion to diverse accelerators ensures that developers aren't locked into NVIDIA ecosystems anymore.
llama.cpp Releases
Coding Toolscoding
llama.cpp b12023 fixes shared prefix bug
This release addresses a specific inference artifact where decision prefixes were incorrectly bleeding into output embeddings. By capping shared prefixes at the first output position, llama.cpp ensures that option tokens remain distinct for parent and child tasks. This is a targeted fix for multi-step reasoning or voting workflows rather than a performance boost. It stabilizes complex generation patterns without changing the underlying model architecture.
