From d65aa1e5cb1c8919b0d05fbd031409bfb5d79c6c Mon Sep 17 00:00:00 2001
From: M-Schwab <matthewschwab256@gmail.com>
Date: Sat, 5 Sep 2026 10:46:53 -0700
Subject: [PATCH] llama-cpp: 0.3.0 -> 0.4.0

https://github.com/ggml-org/llama.cpp/releases/tag/v0.4.0

Assisted-by: opencode + Qwen 3.8 28b: unsloth UD-Q4_K_XL
---
 pkgs/by-name/ll/llama-cpp/package.nix | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/pkgs/by-name/ll/llama-cpp/package.nix b/pkgs/by-name/ll/llama-cpp/package.nix
index 3cc77aacd9b0..d5acf81b6431 100644
--- a/pkgs/by-name/ll/llama-cpp/package.nix
+++ b/pkgs/by-name/ll/llama-cpp/package.nix
@@ -47,8 +47,8 @@
 let
   # Upstream reads these from git, which the release tarball does not ship.
   # They are purely informational: `llama-server --version`, `/props`, and the web UI.
-  buildNumber = "10621";
-  buildCommit = "c1d0e7a";
+  buildNumber = "10809";
+  buildCommit = "5266f24";
 
   # It's necessary to consistently use backendStdenv when building with CUDA support,
   # otherwise we get libstdc++ errors downstream.
@@ -84,7 +84,7 @@ let
 in
 effectiveStdenv.mkDerivation (finalAttrs: {
   pname = "llama-cpp";
-  version = "0.3.0";
+  version = "0.4.0";
 
   __structuredAttrs = true;
   strictDeps = true;
@@ -98,7 +98,7 @@ effectiveStdenv.mkDerivation (finalAttrs: {
     owner = "ggml-org";
     repo = "llama.cpp";
     tag = "v${finalAttrs.version}";
-    hash = "sha256-vVq7+eUN6NXZuqm7Jwlr4iFDV1PjNzQ6nK9AR2zvZYM=";
+    hash = "sha256-WImZjO3U9EXZUNP/FMpxo8PaTjQW8X2SBTfGwwFlZIM=";
   };
 
   patches = [ ];