Models & Labs
Llama.cpp b10899 Release Optimizes Vulkan Performance
llama.cpp Releaseshigh confidence
Why it matters
- →Enhances Vulkan performance for specific computational scenarios.
- →Provides more precise performance tuning with updated tile selection.
- →Strengthens llama.cpp's utility for developers working with Vulkan.
Llama.cpp has released its b10899 update, which brings several optimizations to Vulkan performance, especially for small M matrix operations. The update includes improvements to the m=1 mul_mat function and allows split_k with small M, enhancing computational efficiency. Additionally, the tile selection for coopmat2 now depends on M, not just N, providing more precise performance adjustments. These changes make llama.cpp a more efficient tool for developers using Vulkan and other supported platforms.
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 b10058 release enhances Vulkan support
llama.cpp Releases · July 19, 2026 · Same story
- 2llama.cpp b10831 release enhances Vulkan support
llama.cpp Releases · September 7, 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.
