Llama 3 8b vs GPT-3.5 vs GPT-4 vs Claude Sonnet

2 min readApr 22, 2024

1st Round: Poetry

Prompt: write a rhyming poem about the future of AI

2nd Round: Simple Math

prompt: 72 + 160 / (4 * 4) + 12

3rd Round: Facts

prompt: List the last ten US presidents First and Last names followed by years of presidency

4th Round: Pattern Recognition

prompt: If fish were to be classified according to 5 major categories, list these categories and for each category list 5 fish that are representative of this category


It’s not easy to be the judge, since all got the math correctly, all made great poetry without missing a single rhyme, all got the presidents correctly with gpt 3.5 missing Biden as it is not up to date and last but not least, all creatively divided fish into categories that made sense with GPT-4 classifying biologically when all others based on habitat.

That being said, Llama is the winner since it is much smaller than the rest of the models and is opensource.

What are your thoughts?




Heptaglot Artist, Data Scientist, Filmmaker exploring Creative AI. Started the GAN AI Art Movement (2016). Former Postdoc @CNRS PhD @INFORMATICS. 3xTEDx Speaker