llama.cpp
llama.cpp Update: CPU LLM Inference 3-7x Faster, Quietly Reshaping Enterprise AI
llama.cpp merges a CPU optimization that speeds up matrix ops 3-7x — giving enterprises a viable alternative to the GPU stack they thought they had to
Back to home
open-source-AI
2 articles tagged with this topic
dots-studio
280B Model, 16B Active: Open-Source AI Shifts From Total to Active Parameters
dots3-note preview: 280B MoE, 16B active, 512K context, multimodal. Open-source AI race now measures compute cost, not parameter count.