Download Latest Version llama-b9627-bin-ubuntu-openvino-2026.0-x64.tar.gz (13.9 MB)
Email in envelope

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

Home / b9626
Name Modified Size InfoDownloads / Week
Parent folder
llama-b9626-xcframework.zip < 16 hours ago 206.2 MB
llama-b9626-ui.tar.gz < 16 hours ago 2.7 MB
llama-b9626-bin-win-vulkan-x64.zip < 16 hours ago 38.5 MB
llama-b9626-bin-win-opencl-adreno-arm64.zip < 16 hours ago 11.4 MB
llama-b9626-bin-win-sycl-x64.zip < 16 hours ago 113.8 MB
llama-b9626-bin-win-hip-radeon-x64.zip < 16 hours ago 321.1 MB
llama-b9626-bin-win-cuda-13.3-x64.zip < 16 hours ago 159.0 MB
llama-b9626-bin-win-cuda-12.4-x64.zip < 16 hours ago 261.2 MB
llama-b9626-bin-win-cpu-x64.zip < 16 hours ago 16.9 MB
llama-b9626-bin-win-cpu-arm64.zip < 16 hours ago 10.8 MB
llama-b9626-bin-ubuntu-x64.tar.gz < 16 hours ago 15.5 MB
llama-b9626-bin-ubuntu-vulkan-x64.tar.gz < 16 hours ago 38.4 MB
llama-b9626-bin-ubuntu-vulkan-arm64.tar.gz < 16 hours ago 31.8 MB
llama-b9626-bin-ubuntu-sycl-fp32-x64.tar.gz < 16 hours ago 47.3 MB
llama-b9626-bin-ubuntu-sycl-fp16-x64.tar.gz < 16 hours ago 47.4 MB
llama-b9626-bin-ubuntu-s390x.tar.gz < 16 hours ago 14.5 MB
llama-b9626-bin-ubuntu-rocm-7.2-x64.tar.gz < 16 hours ago 131.1 MB
llama-b9626-bin-ubuntu-openvino-2026.0-x64.tar.gz < 16 hours ago 13.9 MB
llama-b9626-bin-ubuntu-arm64.tar.gz < 16 hours ago 12.5 MB
llama-b9626-bin-macos-x64.tar.gz < 16 hours ago 10.9 MB
llama-b9626-bin-macos-arm64.tar.gz < 16 hours ago 10.6 MB
llama-b9626-bin-android-arm64.tar.gz < 16 hours ago 75.5 MB
cudart-llama-bin-win-cuda-13.3-x64.zip < 16 hours ago 391.0 MB
cudart-llama-bin-win-cuda-12.4-x64.zip < 16 hours ago 391.4 MB
b9626 source code.tar.gz < 16 hours ago 34.9 MB
b9626 source code.zip < 16 hours ago 36.3 MB
README.md < 16 hours ago 4.8 kB
Totals: 27 Items   2.4 GB 0
Add arch support for cohere2-MoE (#24260) * Add arch support for cohere2-MoE * Removed redundant gating_func checks * Changed ffn lookup to prefer prefix_dense_intermediate_size * Renamed arch to cohere2moe * Removed redundant lmhead check and chat template changes * Removed lm_head.weight check from modify tensors, load output tensor not required, fallback to token_embd.weight * Changed to (routed+shared)*0.5 for shared expert combined avg * fixed sliding_window_pattern issue and pattern * Fixed transformers crash 'first_k_dense_replace' error * Remove comment * Removed cohere2-moe as a tokenizer type and kept as tiny_aya. Renamed North-Mini-Code-1.0. * Fixed MTP fail, changed to use iSWA * Fixed remaining todos: cohere2moe renamed, changed swa parsing to use get_key_or_arr, removed extra get_arr use * Force metadata usage Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com> * Remove Cohere2 checkpoint comment Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com> * Remove MTP comment Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com> * Regenerate cohere2moe tokenizer hash * Add cohere2moe to Llama Model Saver supported list * Check for zerobios tensors and add support for Command to use LayerNorm * Map expert_selection_fn to sigmoid in base.py instead of command.py * use bools for foundnorm/foundnormrms Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com> --------- Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

macOS/iOS:

Linux:

Android:

Windows:

openEuler:

  • DISABLED
  • openEuler x86 (310p)
  • openEuler x86 (910b, ACL Graph)
  • openEuler aarch64 (310p)
  • openEuler aarch64 (910b, ACL Graph)

UI:

Source: README.md, updated 2026-06-13