llama.cpp Releases b9297, Adds NVFP4 MTP Scale Tensors

Core summary

One sentence to understand this update

The llama.cpp project released version b9297, primarily adding NVFP4 MTP scale tensors to models and improving macOS/iOS support.

Impact & opportunity

What this could mean

Developers can leverage NVFP4 MTP scale tensors for optimized model performance and efficiency, especially on Apple Silicon, enabling more efficient local LLM inference.