I Tested Claude’s New Visual Tools on ChatGPT.
Claude has always been focused on text and code, but a few days ago it introduced image generation capabilities—ways to produce graphics, layouts, slideshows, and other similar materials created using AI. The toolset is called Claude Design .
Then ChatGPT, having already significantly expanded its investment in AI-powered image generation, released a significant update called Images 2.0 . It promised “quality improvements” in accuracy, consistency, and instruction adherence.
So, what can we do with these AI tools now that wasn’t possible before? And how do their capabilities compare?
Claude Design specializes in business and entrepreneurship.
Claude Design is available to Claude subscribers and, according to Anthropic, is designed for “visual work such as designs, prototypes, slides, one-page presentations, and more.” It won’t create photos of cats skateboarding for you, but it can put together a project slideshow or an iPhone app mockup.
This continues Claude’s trend toward business and entrepreneurship: in addition to programming, teams can use Claude to develop prototypes and create presentations. This may sound a bit boring, but the new features are actually quite versatile and include rotating interactive globes, as shown in the demo video .
To get started, you can use Claude to create visualizations based on existing code, upload existing images and documents as starting points, or simply enter a text prompt. You’ll need to navigate to a dedicated Claude Design landing page , separate from the main chatbot interface, where you can choose how your workflow will work.
I decided to create a presentation showcasing Lifehacker’s value and provided the AI with a few screenshots so it could evaluate the style. Claude then asked me a few questions about my vision, including the combination of text and images, as well as the presentation’s length, before getting started. You can watch the AI ”think” and go through the graphic creation stages in real time.
When the finished work was presented, it looked impressively polished—and Claude Design provides everything you need to export your work elsewhere. One of the most useful features is the ability to customize visual elements after they’re created—in my presentation, I was able to change the accent color, fonts, and slide density with just a few clicks.
You can also request changes using additional prompts and even draw on images to indicate what should happen next. Moving on to the iPhone mockup of a potential Lifehacker news app, Claude Design also did a great job, with the exception of a couple of minor graphic flaws: the app design looked very Lifehacker-esque, and I could request changes simply by drawing on images and typing what I wanted to change.
Everything looks stylish, professional, and easy to use; I can see many companies using Claude Design alongside Claude Code. For individuals, it seems like a useful way to combine design ideas for almost anything, including presentations (though AI tools in apps like PowerPoint and Google Slides might be a better fit).
ChatGPT Images 2.0 is aimed at both consumer and commercial use.
As for ChatGPT and its Images 2.0 update, it has become much more versatile and is aimed at both consumer and enterprise markets. OpenAI claims that suggestions are now more accurate, final results are more precise and consistent, and text rendering has been further improved . Tasks can be more complex, and images appear more “thoughtful.”
Creating images is still incredibly simple: just click the “Create Image” button and explain what you want to see in the prompt. People have created complex “Where’s Waldo?”-style images , infographics based on scientific papers , and cartoonish magazine covers ; I was able to quickly create a Lifehacker comic strip in minutes.
ChatGPT was also able to create two Lifehacker magazine covers of varying quality: they look fairly realistic and are free of obvious errors or typos, yet they also possess the same formulaic quality characteristic of many AI-generated images. Clearly, these covers represent the “average” of all the magazine covers used in ChatGPT’s training data.
In ChatGPT Images 2.0, you won’t be able to create presentations or anything as complex as in Claude Design. While you can theoretically create individual slides and app mockups, there are far more limitations in terms of consistency and subsequent editing of what appears on the screen—ChatGPT is more geared toward one-off, AI-powered experiences.
The new Images 2.0 model is also much better at extracting real information from the internet, so you can create a cartoon map of Middle-earth (though it’s fairly copyright-conscious to avoid copying Tolkien’s work) or create an informative chart about upcoming sports tournaments. I ran both of these tests, and ChatGPT delivered impressive visual results.
Instead of ChatGPT, the closest equivalent to Claude Design is probably the Gemini AI tools available through Google Docs, Sheets, and Slides, which I’ve written about previously . While you can’t create entire presentations yet, you can design slides based on simple prompts and upload other assets as design inspiration.
As in Claude Design, you can use additional hints to refine specific aspects of your slideshow or document design without starting over. And your finished work can be exported to various formats, including PDF and Microsoft Office-compatible files.