llama.cpp Adds Tencent Hunyuan 3 (Hy3) Model Support

Core summary

One sentence to understand this update

llama.cpp updated to b9993, adding support for Tencent Hunyuan 3 (Hy3) model architecture, including MoE decoder stack and MTP speculative decoding.

Impact & opportunity

What this could mean

Builders can now leverage Tencent's advanced Hy3 MoE model within llama.cpp, potentially improving performance and model diversity for local LLM applications.