alt Hacker NewsComparison samples with GPT-Image-2.5 for UI design work:
Starship landing page
Krea 3: https://image.non.io/80b85aaa-286b-4ad7-b3ae-cfab4b491216.we...
GPT-Image-2.5: https://image.non.io/19c5f032-cbc7-4034-9682-3aba451bd669.we...
Fashion history website
Krea 3: https://image.non.io/a6e61219-963c-4921-b453-738960e03308.we...
GPT-Image-2.5: https://image.non.io/c067fd76-9c7e-46c2-a8fd-0b0931bf93f6.we...
Banking website
Krea 3: https://image.non.io/3a1e7b50-36e7-418c-ab47-368ce4f01948.we...
GPT-Image-2.5: https://image.non.io/b96a7e00-3289-4044-90de-e0f86a2ca55e.we...
Overall Krea 3 isn't near GPT-Image-2.5 in terms of visual creativity in this space, but it still should be commended for its ability to render small text details. That said despite it specifying you can add exact hex codes as color reference, I've found that they tend to leak into the design (check the valuation panel of the banking website - notice to the right of the "fixed indexed 7-yr" there's a hex code that made it into the image).
I'd love to get my hands on this to post-train and see if it can be pushed further in this space. I actually see a lot of potential for the model, despite the clear differences in the examples above.
Replies
Thank you. This shows that Krea 3 website isn't being honest, and this is what I hate about all those AI tools: the cherrypicking. Stop wasting our time!
And honestly, the whole thing looks like an ad-for-an-ad.
Thank you. This shows that Krea 3 website isn't being honest, and this is what I hate about all those AI tools: the cherrypicking. Stop wasting our time!
Not sure that UI design is necessarily the ideal or even recommended use case for image models.
The NCC-1701-D really is the default starship in our culture and the training corpus, isn't it.
At least Krea doesn't have the GPT-image shimmer! All of the GPT-image images are immediately recognizable as AI.
oh yea, we are working on enabling people to fine-tune it
yes on gpt-image 2.5 being great at times, at almost 1 order of magnitude more of the cost per image and way slower tho
EDIT: do you find that a lot of people use diffusion models for UI mockups? i’ve always been surprised that people wouldn’t use LLMs exclusively for that instead
That Banking website is a good example of GPT-Image-2.5 having absolutely no self control, to a point I am just shocked they havent been working to fix it. It's like the first thing the model does is "generate 100 slogans" and then try and figure out how to cram them all into the image.
You see it constantly, and its worse with GPT than Gemini models, but they both do it. They seem to think they need to tell not show. They cant not annotate their thoughts into the image. A lot of the time you see either random signs on posts added, tshirts with text describing the story, or hand written annotation over the image.
There's really no reason this should be a problem anymore on default settings, or that everybody should still be getting the exact same style of edgy poster out of GPT with icons for bullets, and lists that add nothing jammed into the image. They need to have some kind of seed library of templates they start from, and higher temperature on selecting different styles.