Download Latest Version llama-b8641-bin-ubuntu-openvino-2026.0-x64.tar.gz (77.0 MB)
Email in envelope

Get an email when there's a new version of llama.cpp

Home / b8638
Name Modified Size InfoDownloads / Week
Parent folder
llama-b8638-xcframework.zip < 12 hours ago 175.8 MB
0
llama-b8638-bin-win-vulkan-x64.zip < 12 hours ago 56.3 MB
0
llama-b8638-bin-win-sycl-x64.zip < 12 hours ago 135.1 MB
0
llama-b8638-bin-win-opencl-adreno-arm64.zip < 12 hours ago 33.1 MB
0
llama-b8638-bin-win-hip-radeon-x64.zip < 12 hours ago 360.1 MB
0
llama-b8638-bin-win-cuda-13.1-x64.zip < 12 hours ago 167.7 MB
0
llama-b8638-bin-win-cuda-12.4-x64.zip < 12 hours ago 249.2 MB
0
llama-b8638-bin-win-cpu-x64.zip < 12 hours ago 39.1 MB
0
llama-b8638-bin-win-cpu-arm64.zip < 12 hours ago 32.0 MB
0
llama-b8638-bin-ubuntu-x64.tar.gz < 12 hours ago 31.6 MB
0
llama-b8638-bin-ubuntu-vulkan-x64.tar.gz < 12 hours ago 48.7 MB
0
llama-b8638-bin-ubuntu-vulkan-arm64.tar.gz < 12 hours ago 40.9 MB
0
llama-b8638-bin-ubuntu-s390x.tar.gz < 12 hours ago 35.2 MB
0
llama-b8638-bin-ubuntu-rocm-7.2-x64.tar.gz < 12 hours ago 159.2 MB
0
llama-b8638-bin-ubuntu-openvino-2026.0-x64.tar.gz < 12 hours ago 76.2 MB
0
llama-b8638-bin-ubuntu-arm64.tar.gz < 12 hours ago 27.8 MB
0
llama-b8638-bin-macos-x64.tar.gz < 12 hours ago 102.3 MB
0
llama-b8638-bin-macos-arm64.tar.gz < 12 hours ago 40.2 MB
0
llama-b8638-bin-910b-openEuler-x86-aclgraph.tar.gz < 12 hours ago 71.8 MB
0
llama-b8638-bin-910b-openEuler-aarch64-aclgraph.tar.gz < 12 hours ago 64.2 MB
0
llama-b8638-bin-310p-openEuler-x86.tar.gz < 12 hours ago 71.8 MB
0
llama-b8638-bin-310p-openEuler-aarch64.tar.gz < 12 hours ago 64.2 MB
0
cudart-llama-bin-win-cuda-13.1-x64.zip < 12 hours ago 402.6 MB
0
cudart-llama-bin-win-cuda-12.4-x64.zip < 12 hours ago 391.4 MB
0
b8638 source code.tar.gz < 15 hours ago 29.6 MB
0
b8638 source code.zip < 15 hours ago 30.8 MB
0
README.md < 15 hours ago 3.4 kB
0
Totals: 27 Items   2.9 GB 0
tests: allow exporting graph ops from HF file without downloading weights (#21182) * tests: allow exporting graph ops from HF file without downloading weights * use unique_ptr for llama_context in HF metadata case * fix missing non-required tensors falling back to type f32 * use unique pointers where possible * use no_alloc instead of fixing f32 fallback * fix missing space

macOS/iOS:

Linux:

Windows:

openEuler:

Source: README.md, updated 2026-04-02