These models work well with http://github.com/Atomic-Germ/Guanaco -- a disk-streaming MoE focused llama.cpp built with patches to upstream main.
-
Atomic-Germ/Qwen3.8-35B-A3B-Distill-GGUF
36B • Updated • 1.5k -
Atomic-Germ/Qwen3.8-Flash-Next-APEX-GGUF
Image-Text-to-Text • 177B • Updated • 3.82k • 5 -
Atomic-Germ/Ornith-1.5-35B-A3B-REAP50-GGUF
Text Generation • 19B • Updated • 2.96k -
Atomic-Germ/Qwen3.8-Whittle-MoE-27B-A17.8B-GGUF
27B • Updated • 1.36k
