diff --git a/misc/llama-cpp/Makefile b/misc/llama-cpp/Makefile
index e91ab62311c4..7939399b16af 100644
--- a/misc/llama-cpp/Makefile
+++ b/misc/llama-cpp/Makefile
@@ -1,40 +1,40 @@
 PORTNAME=	llama-cpp
 DISTVERSIONPREFIX=	b
-DISTVERSION=	3408
+DISTVERSION=	3428
 CATEGORIES=	misc # machine-learning
 
 MAINTAINER=	yuri@FreeBSD.org
 COMMENT=	Facebook's LLaMA model in C/C++ # '
 WWW=		https://github.com/ggerganov/llama.cpp
 
 LICENSE=	MIT
 LICENSE_FILE=	${WRKSRC}/LICENSE
 
 BROKEN_armv7=	clang crashes, see https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=278810
 
 USES=		cmake:testing compiler:c++11-lang python:run shebangfix
 USE_LDCONFIG=	yes
 
 USE_GITHUB=	yes
 GH_ACCOUNT=	ggerganov
 GH_PROJECT=	llama.cpp
 GH_TUPLE=	nomic-ai:kompute:4565194:kompute/kompute
 
 SHEBANG_GLOB=	*.py
 
 CMAKE_ON=	BUILD_SHARED_LIBS
 CMAKE_OFF=	LLAMA_BUILD_TESTS
 CMAKE_TESTING_ON=	LLAMA_BUILD_TESTS
 
 LDFLAGS+=	-pthread
 
 OPTIONS_DEFINE=		EXAMPLES
 OPTIONS_SUB=		yes
 
 EXAMPLES_CMAKE_BOOL=	LLAMA_BUILD_EXAMPLES
 
 BINARY_ALIAS=	git=false
 
 # 1 test fails due to a missing model file (stories260K.gguf)
 
 .include <bsd.port.mk>
diff --git a/misc/llama-cpp/distinfo b/misc/llama-cpp/distinfo
index 8551588a9beb..2820e1fe7e1b 100644
--- a/misc/llama-cpp/distinfo
+++ b/misc/llama-cpp/distinfo
@@ -1,5 +1,5 @@
-TIMESTAMP = 1721307477
-SHA256 (ggerganov-llama.cpp-b3408_GH0.tar.gz) = 12339459fd2d5e1c96814ca8645cea298fd6d13e6cd5b163ca55bd45bae16097
-SIZE (ggerganov-llama.cpp-b3408_GH0.tar.gz) = 20060356
+TIMESTAMP = 1721549594
+SHA256 (ggerganov-llama.cpp-b3428_GH0.tar.gz) = 54464eb17ff2e710c10266b8644c5ef6b5b26016b90705c0eda3bae00d421436
+SIZE (ggerganov-llama.cpp-b3428_GH0.tar.gz) = 20069958
 SHA256 (nomic-ai-kompute-4565194_GH0.tar.gz) = 95b52d2f0514c5201c7838348a9c3c9e60902ea3c6c9aa862193a212150b2bfc
 SIZE (nomic-ai-kompute-4565194_GH0.tar.gz) = 13540496
diff --git a/misc/llama-cpp/pkg-plist b/misc/llama-cpp/pkg-plist
index 4f3dbf036296..c386202d9b63 100644
--- a/misc/llama-cpp/pkg-plist
+++ b/misc/llama-cpp/pkg-plist
@@ -1,48 +1,55 @@
 %%EXAMPLES%%bin/convert_hf_to_gguf.py
 %%EXAMPLES%%bin/llama-baby-llama
 %%EXAMPLES%%bin/llama-batched
 %%EXAMPLES%%bin/llama-batched-bench
 %%EXAMPLES%%bin/llama-bench
 %%EXAMPLES%%bin/llama-bench-matmult
 %%EXAMPLES%%bin/llama-cli
 %%EXAMPLES%%bin/llama-convert-llama2c-to-ggml
 %%EXAMPLES%%bin/llama-cvector-generator
 %%EXAMPLES%%bin/llama-embedding
 %%EXAMPLES%%bin/llama-eval-callback
 %%EXAMPLES%%bin/llama-export-lora
 %%EXAMPLES%%bin/llama-finetune
 %%EXAMPLES%%bin/llama-gbnf-validator
 %%EXAMPLES%%bin/llama-gguf
 %%EXAMPLES%%bin/llama-gguf-hash
 %%EXAMPLES%%bin/llama-gguf-split
 %%EXAMPLES%%bin/llama-gritlm
 %%EXAMPLES%%bin/llama-imatrix
 %%EXAMPLES%%bin/llama-infill
 %%EXAMPLES%%bin/llama-llava-cli
 %%EXAMPLES%%bin/llama-lookahead
 %%EXAMPLES%%bin/llama-lookup
 %%EXAMPLES%%bin/llama-lookup-create
 %%EXAMPLES%%bin/llama-lookup-merge
 %%EXAMPLES%%bin/llama-lookup-stats
 %%EXAMPLES%%bin/llama-parallel
 %%EXAMPLES%%bin/llama-passkey
 %%EXAMPLES%%bin/llama-perplexity
 %%EXAMPLES%%bin/llama-quantize
 %%EXAMPLES%%bin/llama-quantize-stats
 %%EXAMPLES%%bin/llama-retrieval
 %%EXAMPLES%%bin/llama-save-load-state
 %%EXAMPLES%%bin/llama-server
 %%EXAMPLES%%bin/llama-simple
 %%EXAMPLES%%bin/llama-speculative
 %%EXAMPLES%%bin/llama-tokenize
 %%EXAMPLES%%bin/llama-train-text-from-scratch
 include/ggml-alloc.h
 include/ggml-backend.h
+include/ggml-blas.h
+include/ggml-cuda.h
+include/ggml-kompute.h
+include/ggml-metal.h
+include/ggml-rpc.h
+include/ggml-sycl.h
+include/ggml-vulkan.h
 include/ggml.h
 include/llama.h
 lib/cmake/llama/llama-config.cmake
 lib/cmake/llama/llama-version.cmake
 lib/libggml.so
 lib/libllama.so
 %%EXAMPLES%%lib/libllava_shared.so
 libdata/pkgconfig/llama.pc