# llama.cpp 0.6.0 update adds DeepSeek V4 support and new GPU backends

> The open‑source llama.cpp library released version b11890 on Oct 26 2026, adding DeepSeek V4 model fixes and expanded GPU support for CUDA 12/13, Vulkan and more.

Oossa · 2026-10-26 · https://oossa.com/en/llama-cpp-0-6-0-update-adds-deepseek-v4-support-and-new-gpu-backends

The ggml‑org team released llama.cpp version b11890 on Oct 26 2026. The update cleans up DeepSeek V4 code, moves some model parameters into a central config, and drops unused model classes. It also expands hardware support, adding CUDA 12 and 13 libraries, Vulkan, ROCm and OpenVINO backends for Linux, Windows and macOS. The changes let developers run larger models faster on more GPUs and NPU‑enabled devices.

## The facts

- Release date: Oct 26 2026 ("Mon Oct 26 2026 09:54:15 GMT+0100")
- Adds CUDA 12 (12.8) and CUDA 13 (13.4) library support

## Why it matters

Developers can now run DeepSeek V4 models on a wider range of GPUs, reducing the need for cloud‑based inference.

## Sources & references

1. [ggml-org/llama.cpp b11890](https://github.com/ggml-org/llama.cpp/releases/tag/b11890) – llama.cpp, 2026-10-26

Last updated: 2026-10-26