The ggml‑org team released llama.cpp version b11769. The update fixes a bug where the server router did not fully pass client disconnect signals to child processes, which could delay request cancellation. The change also removes a timing‑sensitive test. New pre‑built binaries are now available for macOS Apple Silicon, macOS Intel, iOS, and several Ubuntu variants.
Why it matters
Developers can run llama.cpp locally with more reliable request cancellation, reducing wasted compute on interrupted tasks.
