This website uses cookies

Read our Privacy policy and Terms of use for more information.

Remember those dark, primitive days when AI image tools were spectacularly bad at spelling? We're talking restaurant menus full of gibberish, street signs with scrambled letters, and enough chaotic typos to make your high school English teacher weep.

Well, good news! Those dark ages are officially over.

OpenAI dropped ChatGPT Images 2.0 on Tuesday, April 21, and the headline feature is genuinely wild: the model can finally render clean, legible, and accurate text inside its images.

Ask it for a professional menu, a polished poster, or a product ad, and you’ll get something you could actually use without a second glance. 

ChatGPT Images 2.0

But readable text is just the beginning. Images 2.0 comes packed with what OpenAI calls "thinking capabilities." When you select the Thinking models, this thing doesn't just blurt out an image. It browses the web, plans out the layout, generates multiple distinct images from a single prompt, and—wait for it—proofreads its own work before handing you the final result. 

But that’s not all: OpenAI is promising "unprecedented specificity and fidelity." We’re talking:

  • Small text and UI elements that you can actually read.

  • Dense layouts that aren't a jumbled mess.

  • 2K resolution that looks sharp on any screen.

The model is also making massive strides in multilingual support. It is finally rendering non-Latin scripts, including Japanese, Korean, Chinese, Hindi, and Bengali, with actual accuracy. 

  • Who gets it: All ChatGPT and Codex users have access starting now.

  • The "Pro" Perks: Paid subscribers unlock the more advanced, "thinking" outputs.

  • For the Builders: Developers can tap into the gpt-image-2 API, with pricing that scales based on resolution and quality.

One last piece of gossip? OpenAI is staying tight-lipped about exactly what model architecture is powering Images 2.0 under the hood. But hey, a little mystery never hurts anybody, right?

So are you ready to retire your design software, or do you still have a soft spot for doing things the old-fashioned way? I’m dying to know if this replaces your current creative workflow!

Reply

Avatar

or to participate

More From The Automated