Resolution for custom GPTs stalling and flailing?

Like many other users have posted, customGPT performance has degraded significantly - from ~90 to 95% to 30 to 50%. By any measure it went from great to pitiful. This is certainly not an issue of capability because they were capable not long ago. OpenAI has made them worse either intentionally or inadvertently. Is there either a resolution to fix or an explanation of why it’s not going to be fixed?

1 Like