# llama.cpp b11809 adds WebGPU support for Q2_0

> The llama.cpp release adds Q2_0 row dequantization and matrix multiplication to its WebGPU backend. The release notes give no performance figures.

Oossa · 2026-10-22 · https://oossa.com/en/llama-cpp-b11809-adds-webgpu-support-for-q2-0

The ggml-org team has released llama.cpp b11809, adding support for Q2_0 to its WebGPU backend. The change covers both row dequantization and matrix multiplication, the release notes say. They provide no speed, hardware-compatibility or memory-use results, so the practical performance impact is unclear.

## The facts

- Release: llama.cpp b11809.
- WebGPU change: Q2_0 row dequantization and matrix multiplication.

## Why it matters

Developers using llama.cpp with WebGPU can try Q2_0 models, but the release notes do not say how well they will run on specific devices.

## Sources & references

1. [ggml-org/llama.cpp b11809](https://github.com/ggml-org/llama.cpp/releases/tag/b11809) – llama.cpp, 2026-10-22

Last updated: 2026-10-22