Comparing AI Tools for Generating Images with Text

I have been trying out various AI tools recently. I wanted to create images of promotional merchandise with our simulated company name, Riberpublic SLS. The company specializes in advertising, marketing, and promotional merchandise. My students, working within our simulated company, need to create content for social media, our catalog, and our website. I used Leonardo AI, MidJourney, and DALL-E (via ChatGPT) to see which would give the best results. Here’s what I found.

First, I asked ChatGPT to generate several image prompts because I’m not very good at it, and I find that this tool creates quite accurate prompts:

A realistic image showing a collection of promotional merchandise products, including a mug, T-shirt, notebook, and tote bag, all branded with the logo of ‘Riberpublic SLS’. The logo should be clearly visible on each product, professionally placed, and designed to stand out with vibrant colors. The setting should be on a wooden table with a natural, well-lit environment, making the products look polished and ready for marketing

DALL-E with ChatGPT:

DALL-E produced fairly coherent images, but the text didn’t come out well. It struggled with this aspect. Despite giving many instructions, I couldn’t get the tool to generate truly realistic images. They always looked like drawings. It’s a fast tool but, at least, integrated into ChatGPT 4, it leaves a lot to be desired. I have a Chat GPT plus paid plan.

MidJourney

MidJourney creates striking images for Riberpublic’s portfolio, but it didn’t manage to incorporate the logo clearly on the products. It might work well for more artistic or creative projects, but not for specific branding needs like ours. I have a monthly basic plan to test the tool. This plan allowed me to try generating the images with an actual image of Riberpublic’s logo. That didn’t work well either. Still, I’d confidently use MidJourney for creating faces, where it truly excels.

Leonardo AI: The Best by Far!

Without a doubt, Leonardo AI took the top spot. With the same prompt, it was the only one that managed to generate a realistic image with the text clearly visible on all the products. It’s not perfect, but it’s quite precise, and best of all, it’s free (you get 150 coins a day to generate images, so it is perfect for my students).

After testing all three tools, I’m clear that if you need precise images with text, Leonardo AI is the way to go. For more creative projects, MidJourney would be my favorite, but in our case, Leonardo won hands down.