llama.cpp b12241 fixes FastRPC handle truncation on Hexagon

Posted Nov 12, 07:02 As of Nov 10, 2026

The ggml-org/llama.cpp b12241 release is a minor update that fixes a bug where 64-bit FastRPC handles were truncated on close in Hexagon. The release page does not list new model support or user-facing features.

What happened

Source: ggml-org/llama.cpp GitHub release b12241 (2026-11-10)
Fix: 64-bit FastRPC handle truncation on close in Hexagon (#30952)
Prebuilt binaries: macOS Apple Silicon, Ubuntu, Android Snapdragon (CPU, Adreno GPU, Hexagon NPU), Windows, CUDA, ROCm, Vulkan, SYCL builds
New model support or user-facing features are not stated on the release page
The lazy-grammar fix for tool calls mentioned in a secondary summary was not confirmed on the GitHub page, so it is excluded

Our take

This fix matters for users who run the Hexagon NPU path on Snapdragon devices. The impact on Mac and PC users appears small.
Not confirmed: the effect of this fix on actual speed or stability, and the details of the conditions that trigger it.

Sources and related posts

ggml-org/llama.cpp GitHub 릴리스 b12241 Primary source

Ollama v0.40.16 Stable Release Notes Memory Savings on Apple Silicon (github.com)

Ollama has released the v0.40.16 stable tag. The official release notes state that memory usage for Qwen3.8 Flash Next on Apple silicon has decreased, and they include CUDA 13.5 support for Windows on ARM.

Local AI| Nov 11, 07:02 |1 sources |Our take included

Ollama v0.40.16-rc0: pplx-decider added, API behavior fixed (github.com)

Ollama pre-release v0.40.16-rc0 adds Perplexity's 27B decider model, and changes how the OpenAI-compatible endpoint handles default sampler values. The stable release may differ from these notes.

Local AI| Nov 10, 07:02 |1 sources |Our take included

llama.cpp b12200, Qualcomm OpenCL quantization kernel improvements (github.com)

The llama.cpp b12200 release contains one change that reworks OpenCL q8_0 and small-batch MoE processing. No performance figures were published.

Local AI| Nov 8, 17:43 |1 sources |Our take included