Tried gemma4 on LM Studio

Apr 04, 2026

I also tried gemma4 on LM Studio, but it failed to load the model, not really sure why. See Gibberish results running gemma4 on my integrated GPU on my failed attempt at running gemma4 on ollama.

The Phi3.5 3.8B model worked though. I did a test by setting GPU offloading to 0. You can see the usage of GPU going from 100% to a very low number using intel_gpu_top.

#gemma4 #lm-studio #llm #gpu