Comment by Tenoke

I do wonder how much of this is the prompt. Gpt-4 was doing really well against me when it launched but my prompt was quite different