Maybe I am biased but this seems much better than Dall-E that i use currently
https://openai.com/index/introducing-4o-image-generation/
But … as usual, no info/news on when it will be available on the Azure platform. : Azure OpenAI Service models - Azure OpenAI | Microsoft Learn doesn’t mention anything
Any ideas, anyone?
Thanks
4o image generation rolls out starting today to Plus, Pro, Team, and Free users as the default image generator in ChatGPT, with access coming soon to Enterprise and Edu. It’s also available to use in Sora. For those who hold a special place in their hearts for DALL·E, it can still be accessed through a dedicated DALL·E GPT.
There is also no concrete info when it will come to which accounts. OpenAI rollouts can sometimes last months, so it’s hard to say, unfortunately 
I’d be curious if it can do image+text->image
I guess we’ll see
If anyone gets any information about when 4o Image Generation will be accessible via API, please let us know!
I’m also really excited about image + text → image, as @Diet mentioned 
Yes, it can do Image+Text → Image!
A friend of mine already has access and he made some cool stuff with it already:
I thought the question was about putting text on a given picture.
Here is a not so creative re-work of the ChatGPT interface.
Note that more and more small errors will be introduced when editing the image repeatedly. In this case the ‘My man’s is ChatGPT’ tagline is already showing signs after the first edit.
That’s a pretty neat idea. One where we can start putting gpt-4o’s ability for multimodal I/O to the test.
User’s Task:
A test of vision, accuracy, and understanding: did it miss any when shuffling around words?
Looks good, and double the 512x512 I sent.
A prompt “Click on the first group of four, sending, for example [A3, C1, …] (column, row)” is the next challenge using AI visual understanding of what it made (the second ‘click’ is wrong).
In the new image, the first group of four could be: [A3, A4, B3, C2](“GALLOON”, “THRUST”, “CLOUD”, “DISK”) — possibly themed around movement/force or nautical/space concepts.
The cat is oficially out of the bag - What a time to be alive!
any ideas on how long things typically take to get into Azure?
They’re typically immediate when they get to the API, but there’s often a waiting list.
Yea, I just wanted to write about too (and since there is no official thread announcement here).
It gotten light-years better. From hardly usable to a really great tool. Its finally generating very accurately especially for designs, logos or examples. Hopefully Sora will get there too ))))
So keep that up.
Is this Ghibli style image generation also available via API at the moment … ?
