We also tested on our side and confirmed that the issue has been resolved.
Thank you for taking care of this.
That said, this same type of problem has occurred in the past as well
https://community.openai.com/t/custom-gpt-oauth-stuck-on-talking-to-domain-looking-for-official-guidance/1348110
I understand things can get busy, but we would really appreciate continued efforts to prevent similar regressions in the future.
Thanks again for your support. @OpenAI_Support
We have an enterprise subscription and we’ve noticed with 5.1, the calling of our tools is really much poorer. It’s not calling it and sometimes resorts to public information or weird Python coding, which it tries. Again, it’s a huge step backwards in terms of 5. Therefore, we turned off 5.1 and went back to 5.
Anyone else having trouble with their actions as an enterprise customer?
We’ve put ours on model 4o as the 5 and 5.1 models tend to deviate results more than desired.
There has been a new version release (5.1) that broke my custom GPT and no actions are triggered because of it. The interesting part is that the custom GPT works in the app but in no browsers. Can you pelase check @OpenAI_Support @OpenAIAPIhelper ?
Hello @OpenAI_Support the issue is back and no actions from my custom GPT work. Can you please check and fix? Thanks!
Hi @OpenAI_Support , we’ve also noticed the same behavior. I used to have a custom gpt action that’d send data to a webhook. The gpt would first talk to the connector, than ask me to either confirm or deny, but now the confirm/deny button do no render anymore. It was working at 14:30 PM, but since then none of my employees have been able to use our program