> Gemma 4 is great, but really, Llama, in 2026?

I'd assume the author is just getting confused because of ollama and llama.cpp and all the other ecosystem "llama" that are still in use. Llama really did kick off the open models thing