Any idea when openai will reduce Custom Fine Tune model api inference cost?

Most models’ inference have been drastically reduced in price, other than Fine Tuned models!

When can we expect a price reduction ??

gpt-3.5-turbo fine-tune price was taken from 8x to 3x the base chat model on November 6.

davinci-002 remains at 6x base - but the base model quality of this new model is like downgrading from gpt-3 davinci to gpt-3 babbage… good for finishing a sentence or outputting a token’s statistics if you don’t multi-shot half the context.

The pricing of each and all versions of gpt-3.5-turbo, and all fine-tunes of gpt-3.5-turbo versions, should really be clearly spelled out. Putting the price of only one model out of four, with the others all more expensive, can be deceptive.