Best Gemma 4 GGUFs: Evaluations from Q4 to Q2

APEX or Unsloth's UD GGUFs?

Google released Gemma 4 last week. As usual after a major release, the open-source community moved quickly to produce GGUF versions so people can run the models locally in tools like LM Studio and llama.cpp, with much lower GPU memory requirements.

However, none of these GGUF variants have been properly evaluated yet.