0-suite /
ai0

playground

Two local models, one prompt, side by side — all on your GPU. Models are GB-scale, so the columns run one after the other.

system prompt & temperature
 
 
Because each model is gigabytes, only one can be resident on the GPU at a time — so the two columns run sequentially (column A finishes, then B loads and runs). A fair way to feel the quality/speed trade-off between, say, a 1B and a 3B model. Nothing leaves your device.