# llama.cpp adds fused top‑k mask operation in vb11800 release

> The open‑source inference library now includes a new fused op for QSA masking, improving performance for certain models.

Oossa · 2026-10-21 · https://oossa.com/en/llama-cpp-adds-fused-top-k-mask-operation-in-vb11800-release

The ggml‑org llama.cpp project released version b11800 on 2026-10-21. The update adds a fused top‑k mask operation (ggml_top_k_mask) that can be used by QSA, DSA and DeepSeek V4 CSA models. The op replaces hand‑crafted masking code with a single node that sets non‑selected cells to –inf. It works per‑head and can handle optional scores, block sizes and extra cells.

## The facts

- Release version b11800 published on Wed Oct 21 2026
- New fused op called ggml_top_k_mask for top‑k masking

## Why it matters

Developers can run supported models faster because the library now handles masking in one optimized step.

## Sources & references

1. [ggml-org/llama.cpp b11800](https://github.com/ggml-org/llama.cpp/releases/tag/b11800) – llama.cpp, 2026-10-21

Last updated: 2026-10-21