# llama.cpp release b11352 adds mask optimizations and new binaries

> The ggml‑org/llama.cpp project released version b11352 on Oct 3 2026, improving mask construction for Qwen and GLM5 models and publishing fresh builds for macOS, iOS and Linux.

Oossa · 2026-10-02 · https://oossa.com/en/llama-cpp-release-b11352-adds-mask-optimizations-and-new-binaries

The open‑source LLM library llama.cpp rolled out version b11352. The update speeds up mask building for Qwen‑4‑exp and GLM5‑next models. It also adds pre‑compiled binaries for Apple Silicon, Intel Macs, iOS, and several Linux flavors, including Vulkan and CUDA GPU support.

## The facts

- Release tag b11352 published on Sat Oct 03 2026
- New binaries for macOS Apple Silicon, macOS Intel, iOS, Ubuntu x64/arm64 (CPU, Vulkan, CUDA)

## Why it matters

Developers can run newer LLM models faster on more platforms without building the code themselves.

## Sources & references

1. [ggml-org/llama.cpp b11352](https://github.com/ggml-org/llama.cpp/releases/tag/b11352) – llama.cpp, 2026-10-02
2. [ggml-org/llama.cpp b11355](https://github.com/ggml-org/llama.cpp/releases/tag/b11355) – llama.cpp, 2026-10-02

Last updated: 2026-10-02