@@ -10,6 +10,10 @@ from version 5.0.0 onward. Pre-fork releases (`1.x`–`4.2.0`) were authored by
 ## [Unreleased]
 
 ### Changed
+- **llama.cpp b11968** (from b11960, 8 upstream commits): SVE-256 kernels for the Q6_K gemm and the
+  MXFP4 gemv on ARM CPUs (#27472, #28141), CUDA's MMQ fuses a MoE's gate/up projections with the GLU
+  (#31188), a work-group argmax for short TOP_K rows in SYCL (#31110), DeepSeek-V4's KV cache without
+  the simple batch split (#31300). No source change here.
 - **llama.cpp b11960** (from b11954, 6 upstream commits): llama.cpp 0.8.0 and ggml 0.28.0, tool calls in
   `/v1/realtime` (#31287), video input for LFM3 (#31283), the WebUI's reasoning menu built from the
   template's effort levels (#31273). No source change here.
 
@@ -6,7 +6,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 Java bindings for [llama.cpp](https://github.com/ggerganov/llama.cpp) via JNI, providing a high-level API for LLM inference in Java. The Java layer communicates with a native C++ library through JNI.
 
-Current llama.cpp pinned version: **b11960**
+Current llama.cpp pinned version: **b11968**
 
 ## Natives jars: library jars and GPU module jars (`.github/natives.csv`)
 
@@ -215,7 +215,7 @@ measurements behind the Windows step are in `docs/handover/local-agent-report-b1
   a `llama-common`/`cpp-httplib` that came out shared (one more file to ship). On the static
   platform (macOS) all of it stays inside `libjllama`. **Every ggml default llama.cpp's own
   CMakeLists changes before its `add_subdirectory(ggml)` is repeated** -- `GGML_LLAMAFILE_DEFAULT`
-  and `GGML_CUDA_GRAPHS_DEFAULT` at b11960 (unchanged since b11639) -- because adding ggml first skips that block; re-check with
+  and `GGML_CUDA_GRAPHS_DEFAULT` at b11968 (unchanged since b11639) -- because adding ggml first skips that block; re-check with
   `grep -n _DEFAULT CMakeLists.txt` at the new tag on every bump. The second one was missing once:
   ggml's own default for CUDA graphs is OFF, and a CUDA module built without it lost two thirds of
   its token generation (265.6 -> 88.8 tg128 on an RTX 3070, prompt processing -20%; Vulkan, which
@@ -1233,7 +1233,7 @@ needs no extra step here, `build-webui` re-reads the tag and rebuilds the matchi
 ships no UI):
 ```bash
 # needs node/npm + network for the asset build; the embed step is plain cmake -P
-git clone --depth 1 --branch b11960 https://github.com/ggml-org/llama.cpp /tmp/lc
+git clone --depth 1 --branch b11968 https://github.com/ggml-org/llama.cpp /tmp/lc
 ( cd /tmp/lc/tools/ui && npm ci && npm run build )
 mkdir -p webui-generated /tmp/ui-gen
 cmake -DUI_SOURCE_DIR=/tmp/lc/tools/ui -DUI_BINARY_DIR=/tmp/ui-gen \
@@ -1273,7 +1273,7 @@ cache lives in **Depot Cache** over sccache's **WebDAV** backend:
 - `SCCACHE_WEBDAV_TOKEN: ${{ secrets.DEPOT_TOKEN }}` — a Depot **organization** token, stored
   as the repo secret **`DEPOT_TOKEN`**.
 
-Because `sccache` is **content-addressed** and llama.cpp is pinned (`JLLAMA_GIT_TAG b11960`), the
+Because `sccache` is **content-addressed** and llama.cpp is pinned (`JLLAMA_GIT_TAG b11968`), the
 ~280 upstream object files are byte-identical every run, so a warm cache recompiles only the
 *changed* files. Depot's cache is **shared across all branches** (unlike GitHub's
 per-branch `actions/cache`), so every branch builds incrementally; a `b<nnnn>` version bump
@@ -2868,7 +2868,7 @@ the parameterized `CApiVocabOnly` case expands to 22 instances and `MoeRepack` t
 
 #### Upstream source location (in CMake build tree)
 
-llama.cpp is fetched via CMake FetchContent, pinned to `JLLAMA_GIT_TAG b11960`.
+llama.cpp is fetched via CMake FetchContent, pinned to `JLLAMA_GIT_TAG b11968`.
 
 **GoogleTest** is a separate `BUILD_TESTING`-only FetchContent (`GIT_TAG v1.18.0`), used solely
 by the `jllama_test` C++ unit-test binary — not by the shipped library, and not coupled to the
 
@@ -12,7 +12,7 @@
 **Build:**  
 ![Java 8+](https://img.shields.io/badge/Java-8%2B-informational)  
 ![Platform](https://img.shields.io/badge/Platform-Linux%20%7C%20macOS%20%7C%20Windows%20%7C%20Android-lightgrey)  
-[![llama.cpp b11960](https://img.shields.io/badge/llama.cpp-%23b11960-informational)](https://github.com/ggml-org/llama.cpp/releases/tag/b11960)  
+[![llama.cpp b11968](https://img.shields.io/badge/llama.cpp-%23b11968-informational)](https://github.com/ggml-org/llama.cpp/releases/tag/b11968)  
 [![JPMS](https://img.shields.io/badge/JPMS-modular%20JAR-25A162)](https://openjdk.org/projects/jigsaw/)  
 ![JUnit](https://img.shields.io/badge/tested%20with-JUnit6-25A162)  
 [![JSpecify](https://img.shields.io/badge/JSpecify-1.0.1%20%40NullMarked-25A162)](https://jspecify.dev)  
 
@@ -483,7 +483,9 @@ CLAUDE.md, "Linux s390x".
 
 - [ ] **z17 (`-march=arch15`)** is skipped: clang 18 rejects the flag (clang 19 on cfarm191 too), and
       ggml's configure drops the variant silently. Re-check when the toolchain's clang moves, and
-      measure on a z17 if the farm gets one.
+      measure on a z17 if the farm gets one. Upstream builds `-march=arch15` since b11968 (its zDNN
+      release job) with GCC 15.1 / binutils 2.44 on native Ubuntu 26.04 s390x runners (#31317);
+      which clang accepts the flag is still to check.
 - **Ruled out:** `ubuntu-22.04` as the runner (glibc 2.35 still misses RHEL 8/9, and its cross GCC 11
   cannot build llama.cpp, which needs GCC >= 12). `-static-libstdc++` alone is ruled out too: it
   fixes GLIBCXX but not GLIBC.
 
@@ -929,3 +929,5 @@ Used during `llama.cpp` version bumps: when upgrading, scan this file from the r
 | b11901–b11954 | patches | **All 14 apply unchanged at every tag of the range** (each published tag from b11901 to b11954 checked with the patches applied in order; the fail-loud applier on fresh trees at the five steps; `check-patches.py`: `docs/patches.md` is current; `verify-patches-applied.sh`: 14 applied). Patch-target files in the range: `common/arg.cpp` (`0001`, `0015`; #31242's message and #21231's option, away from both), `tools/server/server.cpp` (`0001`, `0006`, `0007`; #21231's measure role in the dispatch), `tools/server/server-models.cpp` (`0008`; #21231, see its row), `tools/server/server-context.{cpp,h}` (`0002`, `0003`, `0021`; #31271's caps and #31269's realtime slot, away from all three), `ggml/src/ggml-rpc/ggml-rpc.cpp` (`0015`; #31204), `tools/mtmd/mtmd-cli.cpp` and `tools/tts/tts.cpp` (`0001`'s flips). Drop-checks re-read at `f1a7dea99`: `0001` -- `argv = utf8.ptrs.data();` at `common/arg.cpp:1296`, no `common_params_parse_main`; `0002` -- `params_base.load_progress_callback = load_progress_callback;` unguarded at `server-context.cpp:1301`; `0003` -- no `slot_prompt_similarity` in `server-context.h`; `0008` -- no `LLAMA_SERVER_WORKER_CMD` upstream; `0015` -- still no way to stop the RPC server in `ggml-rpc.h`; `0021` -- `slot.lora = params_base.lora_adapters;` at `server-context.cpp:1966` without the clear. The files of `0014`, `0017`, `0018`, `0020`, `0022` and `0023` are not in the range. |
 | b11954–b11960 | `CMakeLists.txt` (llama.cpp 0.8.0, #31309), `ggml/CMakeLists.txt` (ggml 0.28.0), `tools/mtmd/{mtmd.{cpp,h},mtmd-helper.{cpp,h},mtmd-image.{cpp,h},clip.cpp,models/lfm3vl.cpp}`, `tools/server/server-{chat.{cpp,h},realtime.{cpp,h},task.h,decision.cpp}`, `tools/ui/**` | **None needed.** Six commits: the version bumps (llama.cpp 0.7 -> 0.8, ggml 0.27 -> 0.28; this build reads the pair from the fetched tree and leaves its libraries unversioned, so nothing moves), tool calls in `/v1/realtime` (#31287, inside `server-realtime.cpp`, which is compiled in since b11946), video input for LFM3 (#31283: `mtmd_video_config` / `mtmd_get_video_config`, `mtmd_bitmap_set_is_video`, `mtmd_helper` `no_soundtrack`; additive), a Responses-API tool converter in `server-chat.cpp`, and the WebUI's reasoning menu from the template's effort levels (#31273). No new file, no license or lockfile change, request schema and response keys unchanged, the ggml defaults of the root `CMakeLists.txt` unchanged. Built at `8642541b2`: 0 warnings, `ctest` 728/728; `TtsIntegrationTest`, `TtsChatTemplateIntegrationTest` (LFM3-Audio), `MultimodalIntegrationTest`, `ToolCallingIntegrationTest`, the router and attach tests, `NativeLibraryLoadSmokeTest` and `LlamaModelTest` green (98 tests). |
 | b11954–b11960 | patches | **All 14 apply unchanged** (`git apply` in order at `8642541b2`, the fail-loud applier on a fresh tree); no patch-target file is in the range, so the drop-checks of the b11901–b11954 row stand. |
+| b11960–b11968 | `ggml/src/ggml-cpu/arch/arm/repack.cpp`, `ggml/src/ggml-cuda/{ggml-cuda.cu,mmq.cu}`, `ggml/src/ggml-sycl/ggml-sycl.cpp`, `src/llama-kv-cache-dsv4.cpp`, `gguf-py/**`, `tests/**`, `.github/workflows/{build-ibm,release,release-publish}.yml`, `CODEOWNERS` | **None needed.** Eight commits: SVE-256 kernels for the Q6_K gemm (#27472) and the MXFP4 gemv (#28141), the fused MoE gate/up + GLU in CUDA's MMQ (#31188), a SYCL work-group argmax for short TOP_K rows (#31110), DeepSeek-V4's KV cache without the simple batch split (#31300), a gguf-py rounding fix, CODEOWNERS. No new file, no CMake, license, lockfile or WebUI change, request schema and response keys unchanged, the ggml defaults of the root `CMakeLists.txt` unchanged. Noted for the z17 item of `TODO.md`: upstream's s390x CI moved to native Ubuntu 26.04 runners (#31317), and its new zDNN release build compiles `-march=arch15` with GCC 15.1 / binutils 2.44. Built at `ac8e1694d`: 0 warnings, `ctest` 728/728; `LlamaModelTest`, `TtsIntegrationTest`, `MultimodalIntegrationTest`, `ToolCallingIntegrationTest`, `RerankingModelTest`, the router and attach tests, `NativeLibraryLoadSmokeTest` and `LlamaCppVersionTest` green (97 tests). |
+| b11960–b11968 | patches | **All 14 apply unchanged** (the fail-loud applier on a fresh tree at `ac8e1694d`); no patch-target file is in the range, so the drop-checks of the b11901–b11954 row stand. |
@@ -356,8 +356,8 @@ endif()
 # hash, every other consumer keeps the number, and the pin check after FetchContent_MakeAvailable
 # asserts that the fetched commit's count IS the number. Clear it once the tag exists;
 # `llama-next-version.sh` says when, and `llama-next-version.sh --resolve b<nnnn>` prints the hash.
-set(JLLAMA_GIT_TAG    b11960)
-set(JLLAMA_GIT_COMMIT "8642541b255bfb5dc39cd351a178be09013a55d6")   # the commit the tag b11960 names; kept as the pin (a tag name can move)
+set(JLLAMA_GIT_TAG    b11968)
+set(JLLAMA_GIT_COMMIT "ac8e1694d837b11ed0aab944a6a2eee34b9c4d80")   # the commit the tag b11968 names; kept as the pin (a tag name can move)
 if(NOT JLLAMA_GIT_TAG MATCHES "^b[0-9]+$")
     message(FATAL_ERROR "JLLAMA_GIT_TAG must be a llama.cpp release number b<nnnn>, got '${JLLAMA_GIT_TAG}'")
 endif()
 
@@ -22,7 +22,7 @@ texts are not needed here.
 
 | Component | Version | Compiled into | SPDX | Texts | Source |
 |---|---|---|---|---|---|
-| llama.cpp (ggml / llama / mtmd / llama-common / tools/server / CPU+Metal+GPU kernels / WebUI own code) | b11960 8642541b255bfb5dc39cd351a178be09013a55d6 (texts unchanged since fa3645feb) | all 23 jars + both AARs | MIT | `llama.cpp` | https://github.com/ggml-org/llama.cpp/blob/fa3645feb5f47cbb30df5e6d031c22543d43d240/LICENSE |
+| llama.cpp (ggml / llama / mtmd / llama-common / tools/server / CPU+Metal+GPU kernels / WebUI own code) | b11968 ac8e1694d837b11ed0aab944a6a2eee34b9c4d80 (texts unchanged since fa3645feb) | all 23 jars + both AARs | MIT | `llama.cpp` | https://github.com/ggml-org/llama.cpp/blob/fa3645feb5f47cbb30df5e6d031c22543d43d240/LICENSE |
 | nlohmann/json | 3.12.0 + vendor/nlohmann/json-deep-nesting.patch | library jars | MIT | `nlohmann-json` | https://github.com/ggml-org/llama.cpp/blob/fa3645feb5f47cbb30df5e6d031c22543d43d240/licenses/LICENSE-jsonhpp |
 | cpp-httplib | v0.62.0 | library jars | MIT | `cpp-httplib` | https://github.com/ggml-org/llama.cpp/blob/fa3645feb5f47cbb30df5e6d031c22543d43d240/vendor/cpp-httplib/LICENSE |
 | BoringSSL (ssl + crypto; static; OPENSSL_NO_ASM) | 0.20261020.0 = d3bf4795ad5b79d646cddf2fec687da5b9b502db | library jars except cpu-android-* and cpu-linux-s390x | Apache-2.0 | `boringssl` | https://boringssl.googlesource.com/boringssl/+/refs/tags/0.20261020.0/LICENSE |
@@ -47,7 +47,7 @@ texts are not needed here.
 | WebUI vendored big-integer | in tree | library jars (embedded ui.cpp) | Unlicense | `webui-big-integer` | tools/ui/src/lib/vendors/big-integer/LICENSE |
 | WebUI vendored decimal.js | in tree | library jars (embedded ui.cpp) | MIT | `webui-decimal.js` | tools/ui/src/lib/vendors/decimal.js/LICENCE.md |
 | WebUI vendored nerdamer-prime | in tree | library jars (embedded ui.cpp) | MIT | `webui-nerdamer-prime` | tools/ui/src/lib/vendors/nerdamer-prime/LICENSE |
-| WebUI npm bundle (JS + CSS incl. KaTeX fonts + workbox) | tools/ui/package-lock.json at b11960 | library jars (every library build job downloads webui-generated) (upper bound (tree-shaking / server-only deps)) | 337 pkgs: MIT 271, ISC 43, BSD-3-Clause 9, Apache-2.0 5, BSD-2-Clause 2, MPL-2.0 OR Apache-2.0 1, Unlicense 1, 5 without field (MIT files); 19 MIT packages ship no license file, the MIT text follows them | `webui-npm` | webui-npm/packages.csv |
+| WebUI npm bundle (JS + CSS incl. KaTeX fonts + workbox) | tools/ui/package-lock.json at b11968 | library jars (every library build job downloads webui-generated) (upper bound (tree-shaking / server-only deps)) | 337 pkgs: MIT 271, ISC 43, BSD-3-Clause 9, Apache-2.0 5, BSD-2-Clause 2, MPL-2.0 OR Apache-2.0 1, Unlicense 1, 5 without field (MIT files); 19 MIT packages ship no license file, the MIT text follows them | `webui-npm` | webui-npm/packages.csv |
 | WebUI third-party logos (huggingface.ico / exa.ico / context7.png / github-*.png) | - | library jars (yes) | trademarks (no license file) | `(not shipped)` | tools/ui/static |
 | CCCL (CUB / Thrust / libcu++ headers) | v3.4.3 | cuda13-linux-x86-64 + cuda13-windows-x86-64 | Apache-2.0 WITH LLVM-exception AND Apache-2.0 AND BSD-3-Clause (CUB) | `cccl` | https://github.com/NVIDIA/cccl/blob/v3.4.3/LICENSE |
 | CUDA Toolkit headers / libdevice | CUDA 13.4 | cuda13 modules (yes) | LicenseRef-NVIDIA-CUDA-EULA | `(not shipped)` | - |
 
@@ -9,28 +9,28 @@
  * library was compiled against, exposed as a compile-time constant so callers can render a badge or
  * emit a startup log line without loading the native library.
  *
- * <p>{@link #LLAMA_CPP_VERSION} is a pure-Java string ({@code "b11960"}) that mirrors the
+ * <p>{@link #LLAMA_CPP_VERSION} is a pure-Java string ({@code "b11968"}) that mirrors the
  * {@code GIT_TAG} in {@code llama/CMakeLists.txt}. It is available even when {@code libjllama} is
  * absent (pure-Java checkout, before {@code System.load}), which is what makes it suitable for a
  * lightweight version badge in Android or other UIs.</p>
  *
  * <p>For the <em>authoritative</em> value that is baked into the native binary — the build number
- * plus the resolved upstream commit, e.g. {@code "b11960-<commit>"} — call
+ * plus the resolved upstream commit, e.g. {@code "b11968-<commit>"} — call
  * {@link net.ladenthin.llama.LlamaModel#getLlamaCppBuildInfo()} instead; that reads llama.cpp's own
  * {@code build-info} through JNI and therefore cannot drift from the compiled library (but requires
  * the native library to be loaded).</p>
  */
 public final class LlamaCppVersion {
 
     /**
-     * The pinned llama.cpp release tag this library was built against, e.g. {@code "b11960"}.
+     * The pinned llama.cpp release tag this library was built against, e.g. {@code "b11968"}.
      *
      * <p>Kept in lockstep with {@code GIT_TAG} in {@code llama/CMakeLists.txt} — see the
      * "Upgrading/Downgrading llama.cpp Version" checklist in {@code CLAUDE.md}. This is the
      * compile-time pin; use {@link net.ladenthin.llama.LlamaModel#getLlamaCppBuildInfo()} for the
      * value actually linked into the native binary.</p>
      */
-    public static final String LLAMA_CPP_VERSION = "b11960";
+    public static final String LLAMA_CPP_VERSION = "b11968";
 
     // Constants holder — not instantiable.
     private LlamaCppVersion() {}