diff --git a/Dockerfile b/Dockerfile
index 4e8e29cb27a9..fa885dd346d4 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -72,6 +72,12 @@ RUN <<EOT bash
     fi
 EOT
 
+RUN <<EOT bash
+    if [ "${BUILD_TYPE}" = "cublas" ] && [ "${TARGETARCH}" = "arm64" ]; then
+        echo "nvidia-l4t" > /run/localai/capability
+    fi
+EOT
+
 # If we are building with clblas support, we need the libraries for the builds
 RUN if [ "${BUILD_TYPE}" = "clblas" ] && [ "${SKIP_DRIVERS}" = "false" ]; then \
         apt-get update && \
diff --git a/backend/cpp/llama-cpp/Makefile b/backend/cpp/llama-cpp/Makefile
index 7a1652aded0e..cf1c3359c8e4 100644
--- a/backend/cpp/llama-cpp/Makefile
+++ b/backend/cpp/llama-cpp/Makefile
@@ -1,5 +1,5 @@
 
-LLAMA_VERSION?=3f4fc97f1d745f1d5d3c853949503136d419e6de
+LLAMA_VERSION?=c7f3169cd523140a288095f2d79befb20a0b73f4
 LLAMA_REPO?=https://github.com/ggerganov/llama.cpp
 
 CMAKE_ARGS?=
diff --git a/docs/data/version.json b/docs/data/version.json
index d7a882dd467d..ca2b056dd181 100644
--- a/docs/data/version.json
+++ b/docs/data/version.json
@@ -1,3 +1,3 @@
 {
-  "version": "v3.1.1"
+  "version": "v3.2.1"
 }