# llama.cpp adds optional Vulkan device keepalive in new release

> The open‑source llama.cpp library (v b12363) now lets Vulkan back‑ends keep the GPU device alive for a set time, easing repeated inference calls.

Oossa · 2026-11-13 · https://oossa.com/en/llama-cpp-adds-optional-vulkan-device-keepalive-in-new-release

The ggml‑org team released llama.cpp version b12363 on November 13, 2026. The update adds an optional device keepalive setting to the Vulkan backend, letting users specify how long the GPU stays active before shutting down. The change moves ownership of the device to the backend and simplifies the code. Pre‑built binaries for macOS, iOS, and several Ubuntu variants are included.

## The facts

- Release version b12363 published on 2026‑11‑13
- Adds optional Vulkan device keepalive (configurable in seconds)

## Why it matters

Developers can run repeated model calls on GPU without costly re‑initialization, improving response speed for local AI apps.

## Sources & references

1. [ggml-org/llama.cpp b12363](https://github.com/ggml-org/llama.cpp/releases/tag/b12363) – llama.cpp, 2026-11-13

Last updated: 2026-11-13