Comment by arpowers

Could use some examples of the improved output. It shouldnt be that hard.

What are each models strengths? I know GPT4 has a strength with code.