Apples to Apples: MLX vs. Llama.cpp for Gemma 4 12B on an M1 16GB(ziraph.com)
ziraph.com
Apples to Apples: MLX vs. Llama.cpp for Gemma 4 12B on an M1 16GB
https://ziraph.com/blog/apples-to-apples-mlx-vs-llama-cpp-gemma-4
https://ziraph.com/blog/apples-to-apples-mlx-vs-llama-cpp-gemma-4
It would not change the core of my article since the bottleneck remains the memory bandwidth on the old M1 16GB though