one prompt / parallel readout
Compare the same thought across the stack.
A small, honest bench for seeing how configured models interpret one prompt. The prompt stays here; each response gets its own lane.
01
Reading channel list…Set the transmission
0 / 4000 characters
02
Waiting for a transmissionRead the response lanes
Run a comparison to place each model’s answer on the same bench.