Same prompt, Same person in different models

I used the same prompt and trained lora of a given person in Cyberrealistic and Proton, Both provided realistic details where one gave more importance to details such as eye colors to be exact while the other focused on keeping it real