If you think that the average public LLM is equivalent to ChatGPT or GPT-4 then you are completely mistaken. By a factor of say 500-10000%.