# llama.cpp vb12329 adds CUDA weight‑scale fusion and new binaries

> The open‑source LLM runtime releases version b12329 with CUDA optimizations and pre‑built binaries for macOS, iOS and Linux.

Oossa · 2026-11-12 · https://oossa.com/en/llama-cpp-vb12329-adds-cuda-weight-scale-fusion-and-new-binaries

The ggml‑org team released llama.cpp version b12329. It adds a CUDA optimisation that fuses weight‑scale handling into matrix‑multiply writes, speeding up nvfp4 checkpoints. The update also includes new pre‑built binaries for Apple Silicon, Intel macOS, iOS and several Linux flavours. The changes are available for download from the GitHub release page.

## The facts

- Release version: b12329
- Published: 2026-11-12

## Why it matters

Developers can run Llama models faster on NVIDIA GPUs and get ready‑to‑run binaries for more platforms.

## Sources & references

1. [ggml-org/llama.cpp b12329](https://github.com/ggml-org/llama.cpp/releases/tag/b12329) – llama.cpp, 2026-11-12

Last updated: 2026-11-12