llama.cpp Update b11065 Optimizes CUDA for Gemma 4 on Ampere GPUs

Core summary

One sentence to understand this update

llama.cpp has released update b11065, which includes specific CUDA optimizations for Gemma 4 models running on Ampere or newer GPU architectures.

Impact & opportunity

What this could mean

Builders working with Gemma 4 models on compatible NVIDIA GPUs will see improved performance and efficiency for local LLM inference through llama.cpp.