This result really surprised us too, but it seems like we were not alone in observing this: https://medium.com/@gabrielgrinberg/openai-api-fine-tuned-gpt-3-5-vs

This result really surprised us too, but it seems like we were not alone in observing this: https://medium.com/@gabrielgrinberg/openai-api-fine-tuned-gpt-3-5-vs-base-gpt-3-5-a-case-study-f3619b4f8cd8 https://medium.com/@gabrielgrinberg/openai-api-fine-tuned-gp... That being said, both accounts that we experimented on were personal accounts ( with a very limited budget) so maybe there are some special behaviors (e.g. some rate limit that only exists for base models)