When 4o imagegen will be implemented in custom GPTs?

GPT builders
, , , ,
1

When will 4o imagegen be available in custom GPTs?

The image shows a partial view of a person in a green apron with a message about old image generation, stating new images are coming soon. (Captioned by AI)

We urgently need the new 4o image generation model in place of DALL·E in the custom GPT editor.

Right now, any time a GPT is used, it defaults to outdated image generation, which limits quality and user experience.

The newer 4o model should be the default for all custom GPTs to produce for better, more realistic, and faster results..

Please prioritize this upgrade.

3

It just got in there.

The image shows a ChatGPT editor interface with various document files listed on the left and a highlighted "Canvas" capability. (Captioned by AI)
The image shows a ChatGPT editor interface with various document files listed on the left and a highlighted "Canvas" capability. (Captioned by AI)960×1079 79.5 KB

P.S.: BUT, it seems like, additionally right now custom gpts can not properly be saved right now at least changing the older ones and then trying to save fails for me…

4

That’s awesome, thanks for noticing..
I’ve also encountered saving errors sometimes, but usually it works if I try again later.

1 Like
5

Yeah, it seems like a lot is off right now. But many times we experience some glitches. Then it gets better again. :slight_smile:

For example:

  • ChatGPTs refusing to work
  • Can’t be saved
  • “Unexpected error occurred” state messages and the like.
  • Chats broken.
  • Tasks/Prompts not being executed properly.

ALL this I experienced since the last two days. I’d guess there’s another big update coming.

It’s always worthwile having a look here:

The image shows an OpenAI system status page, indicating current issues with Sora and a summary of uptime percentages for various services like APIs and ChatGPT. (Captioned by AI)
The image shows an OpenAI system status page, indicating current issues with Sora and a summary of uptime percentages for various services like APIs and ChatGPT. (Captioned by AI)1916×2158 325 KB

(Screenshot is from RIGHT NOW - when the post was made)

1 Like
6

And by the way:

If your custom GPTs are not working. Or you wanna try out more (without ANY voice mode for now):

You can just copy/put the instructions into a project that you’ll create.
Upload your files there.

If you prepared them in a way so you only need the uploaded files your custom gpt will initialize there. THEN, you can also test it on other models.

Example:

The image shows a computer screen with a project management interface and a person thoughtfully sitting by a window. (Captioned by AI)
The image shows a computer screen with a project management interface and a person thoughtfully sitting by a window. (Captioned by AI)1920×3507 340 KB

P.S.: You’ll lose (for now at least) the ability to speak to this project. BUT, they microphone offers at least a “dictate” option for now.

Rumors tell: OpenAI is experimenting with voice mode in projects, though. Let’s see what happens in a while.

P. S. Put this also into here now:

1 Like