I Tried Chatgpt Images 2.0: A Fun, Huge Leap - And Surprisingly Useful For Real Work

2 hours ago

A fun, immense leap - and amazingly useful for existent work

Follow ZDNET: Add america arsenic a preferred source connected Google.

ZDNET's cardinal takeaways

Images 2.0 delivers meticulous matter and usable graphics.
It tin lucifer marque styles, including ZDNET visuals.
Errors still gaffe in, requiring quality review.

Earlier this week, OpenAI unveiled ChatGPT Images 2.0, its caller image procreation engine. Key to this merchandise is simply a jump successful functionality from creating "decorations" (OpenAI's term) to full-page graphics, including elaborate text.

I had early entree to a pre-release version. It worked rather well, but kept messing up connected nan ZDNET logo. Now that nan merchandise has been officially released, I'm giving it an in-depth trial crossed a wide scope of challenges.

Images 2.0 is disposable to each ChatGPT tiers, but nan much tin connection features are only disposable to paying tiers that tin usage nan Thinking model. I'm moving each these tests utilizing a ChatGPT Plus relationship pinch Thinking turned on.

Also: I put GPT-5.5 done a 10-round test: It scored 93/100, losing points only for exuberance

Let's get started pinch nan ZDNET branding exercises. Rather than conscionable uploading ZDNET pages and having it find nan logo connected nan page, I created a standalone image of nan ZDNET logo and uploaded that pinch each prompt. That seemed to thief tremendously.

[One speedy note: ZDNET doesn't licence OpenAI to scrape its pages. Ziff Davis, ZDNET's genitor company, revenge an April 2025 suit against OpenAI, alleging it infringed Ziff Davis copyrights successful training and operating its AI systems. So I utilized a Chrome hold to seizure full-screen screenshots of nan articles I wanted to trial pinch Images 2.0. That's really ChatGPT was capable to publication them.]

Can Images 2.0 sphere nan ZDNET logo?

My starting constituent was nan article I antecedently wrote astir Images 2.0. I fed ChatGPT this prompt: "Create a elaborate and vivid infographic of this article utilizing nan ZDNET marque style and nan attached ZDNET logo."

Not only is nan logo correct, but nan coloring is cleanable for ZDNET. But wherever nan image really shines is its usage of text. All nan matter is correct, moreover nan mini matter connected an perspective successful nan image.

Can it nutrient styled sketchnotes?

Next, I decided to revisit the sketchnotes challenge I gave to Google's Nano Banana a fewer months ago. The duty astatine that clip was to create a sketchnotes type of nan US Bill of Rights. Nano Banana did a awesome occupation pinch nan images, but I had to effort complete and complete (and over) to person it to get nan wording right. Read nan article to spot nan hoops I had to jump through.

Also: I utilized Nano Banana 2 to make cleanable sketchnotes: 5 lessons learned

For ChatGPT Images 2.0, I upped nan stakes slightly. I wanted sketchnotes, but I wanted them successful ZDNET's branding style. I'm playing up nan branding style passim this article because that's 1 measurement ChatGPT Images 2.0 could supply existent worth to users.

Here's nan first prompt: "Make maine a sketchnote of nan US Bill of Rights. Use nan ZDNET logo style and make nan sketchnotes successful nan ZDNET style." That's nan image connected nan left. Here's nan 2nd prompt: "Include nan ZDNET logo and adhd much neon-style colors, possibly connected a achromatic background." That's nan image connected nan right.

First, announcement that nan matter is correct. There are nary duplicates. Nothing is missing. Already, this is caput and shoulders supra Nano Banana's performance. Both versions fresh pinch ZDNET's style. The only point I'm not thrilled pinch is that nan ZDNET logo looks jammed successful connected nan 2nd image. Even so, nan logo is correct, and I could astir apt do a fewer much punctual passes to get it placed better.

Wacky nosy pinch an infographic

But now we travel to nan unforced correction my testing group revealed. I asked Images 2.0 to person my AI website builder shootout article to an infographic. It produced a reasonably usable, if somewhat busy, infographic. It moreover went to nan net and added accusation I didn't person successful nan article, for illustration guidelines pricing.

But location are 4 clear errors:

The header highlights "here are 9 of nan champion AI website builders." It moreover makes nan "9" guidelines out. Except that only 5 website builders were reviewed. Lower successful nan infographic, it shows nan 5 I do review. Oops.
The services I reviewed were Hostinger, GoDaddy, Wix, 10Web, and Squarespace. ChatGPT decided, for immoderate reason, to switch 10Web pinch Durable (a competitor to 10Web). I didn't reappraisal Durable. I didn't moreover mention Durable. Wacky.
The AI produced a summary array for nan services, listing prima ratings for easiness of use, creation flexibility, and AI features. But I didn't supply prima ratings for these categories. The AI was overly generous toward immoderate vendors, successful a measurement that was straight contrary to nan reappraisal matter itself. Odd.
Finally, and this is simply a nit, but still. Way down astatine nan bottom, wherever nan AI correctly reproduced nan ZDNET logo, there's a drooping statement conscionable supra it. Why?

Also: The champion AI image generators: There's only 1 clear victor now

To beryllium fair, these are each errors an in-house quality schematic designer mightiness nutrient successful a first draft. In my years arsenic a laminitis and a merchandise manager, I've surely seen much egregious graphics errors travel backmost from my designers connected their first drafts.

When I re-prompted Images 2.0 pinch corrections (except for nan prima ratings, which I didn't correct successful nan 2nd image), it did correctly modify nan infographic pinch much due information.

ChatGPT Images has travel a agelong measurement

This Images 2.0 merchandise is simply a immense betterment complete erstwhile versions. The ChatGPT Images type I looked astatine past year was impressive, particularly for recontextualizing images.

Also: I sewage an early look astatine ChatGPT Images 2.0, and it's awesome - pinch 1 exception

This caller version, which tin construe existent contented and past create images, is simply a immense leap complete erstwhile builds. More to nan point, it tin present very tangible business value, which makes it worthy a batch not only for nosy pictures but for existent work.

Stay tuned, because I'll beryllium looking astatine really this build compares pinch Google Gemini's Nano Banana. I'll beryllium pushing it moreover further to spot what different work-related tasks it tin thief with, peculiarly erstwhile it comes to personification interface design.

How comfortable are you relying connected AI-generated visuals, knowing that nan exemplary tin present subtle actual errors? Let america cognize successful nan comments below.

You tin travel my day-to-day task updates connected societal media. Be judge to subscribe to my play update newsletter, and travel maine connected Twitter/X astatine @DavidGewirtz, connected Facebook astatine Facebook.com/DavidGewirtz, connected Instagram astatine Instagram.com/DavidGewirtz, connected Bluesky astatine @DavidGewirtz.com, and connected YouTube astatine YouTube.com/DavidGewirtzTV.