Transcription
Check these infographics. Lots of text. Zero spelling mistakes. Insane details. I made all of these using ChatGPT Images 2.0 and honestly, I'm blown away. But is it really better than Gemini? I spent two days testing both and the results might surprise you. Let me show you.
I've been making Gemini tutorials for months, so I know exactly where it shines and where it falls apart, especially when it comes to text. OpenAI claims the new GPT images 2.0 fixes all that with better text rendering, improved detail, and even multilingual support. These are big claims, so I'm putting both tools head to head across 3 text-heavy scenarios, including one test most AI YouTubers completely ignore.
I'll start the test with a cool prompt I found on Sifu Yik’s substack that creates educational infographics. Link’s in the description. Let's first try ChatGPT. In a new chat, click the plus icon and select Create Image to activate images 2.0. Then paste the prompt shown on your screen. Change whatever is in brackets with what you want to visualize. I'm going with the life of Albert Einstein. When done, click send and wait. The process is pretty slow, so be patient.
The generation is over. I must first download the image to take a closer look. To do so, I'll click here. Let's see it now. Wow, so many details are depicted in the image and I don't see any spelling mistakes at all. Even the tiniest bits of text are perfectly accurate. The accuracy is so high that I'm genuinely impressed.
Now I’ll run the same prompt in Gemini and compare the outputs. I believe you already know the drill. Click ‘tools’, enable ‘create images’ and thinking mode. Then paste the prompt and change what's in brackets. And wait. Gemini is done. That's new. There is now a Gemini warning that the small text annotations might contain mistakes. Let's see. Well, I love the visual details. They actually look better than ChatGPT’s. But on the other hand, there is less text compared to the previous image. Oh, and check this. As expected, there are spelling mistakes. Even more down here. You'll have to fix these manually like I have shown you in the video shown on your screen.
All right, time for another test. We all love sketchnotes, don't we? I even made a video on how to build them. This time, to make sure both tools work with identical data, I'll feed them the exact information I want to be visualized. For that, I'll use a different tool to perform the research. I'll switch to Claude and tell it to identify and explain time management techniques, and also be analytic, specific, and to the point, and provide the relevant examples.
All right. Claude has finished. I'll copy its output and return to ChatGPT in a new conversation. I'll paste the information copied from Claude and tell ChatGPT to just store this information and wait for my command. All right, now I'll turn on ‘Create Image’ and run my signature prompt for creating sketchnotes. As always, you'll find it in the pinned comment below.
Okay, time to check the result. The visuals look solid with plenty of detail packed in. Everything looks spot on with no errors inside. There are calendars, processes, diagrams, and yeah, no issues with the font. ChatGPT nailed it.
Now I'll try making sketchnotes with Gemini and compare. Same process as before. I'll first paste Claude’s output and then run the prompt and see what I'll get. It's done. Let's see it. Visually, everything looks perfect and the colors are more vivid, but I bet if we take a closer look, we'll find spelling mistakes. Oh, and there it is. See? It repeated the same sentence twice. And that's not the only issue. Look at this one below. Gemini misspelled the word “urgent”.
Since many of you are not English speakers, I'm sure you want to know which one performs better with languages other than English. So I'll have them both create handwritten studynotes in Greek, which is my native language. I'll begin with ChatGPT. I'll ask it to research the world's most influential civilizations and share its reply in Greek. All right. I'll now follow up with the second prompt and ask it to turn that information into an educational infographic on edged, spiral bound notebook paper.
All right. It has finished. Again, the result looks flawless. No spelling mistakes whatsoever. The visuals could have been a bit more vivid, but I'll take perfect text over pretty visuals any day.
Now, let's try Gemini. We'll follow the same process as before. And here is the result. Gemini's visuals are slightly better than ChatGPT’s, I'll give it that, but again, it seems there's way less text on the page. My guess is Gemini has played it safe to avoid spelling mistakes, and this time it worked. I can't find any errors whatsoever, but is that really a win when half of the content is missing?
And here comes the moment of truth. I've made dozens of tutorials on using Gemini for generating infographics, timelines, and all kinds of educational materials, but I always noticed issues with text’s accuracy. ChatGPT images 2.0, on the other hand, is performing exceptionally well. Sure, it's slower, but it's way more precise and produces zero spelling errors. And this is huge. So ladies and gentlemen, we have a new king for generating images with text. And this is now ChatGPT. But images are just a fraction of the awesome things you can do with ChatGPT. So if you want to watch The Ultimate Guide to Mastering ChatGPT for Productivity and Business, check out the video shown on the screen. Till the next time. See you.
Time to jump to our third test. Now I'll try making sketchnotes with Gemini and compare. Same process as before. I'll first paste Claude’s output, then run the prompt and see what I'll get. Again, it looks flawless. No spelling mistakes whatsoever. The visuals could have been a bit more vivid, but I'll take perfect text over perfect visuals any day. All right. It has finished.
Check these infographics. Lots of text. Zero spelling mistakes. Insane details. I made all of these using images 2.0, and honestly, I'm blown away. But is it really better than Gemini? I spent two days testing both and the results might surprise you. Let me show you.
I've been making Gemini tutorials for months, so I know exactly where it shines and where it falls apart, especially when it comes to text. OpenAI claims the new GPT images 2.0 fixes all that with better text rendering, improved detail, and even multilingual support. Big claims. So I'm putting both tools head to head across three text-heavy scenarios, including the one most AI completely ignore.
Check these infographics. Lots of text, zero spelling mistakes. And check these infographics. Lots of text, zero spelling mistakes. Insane details. I made all of these using images 2.0 and honestly, I'm blown away. But is it really better than Gemini? I spent two days testing both and the results might surprise you. Let me show you.
I've been making Gemini tutorials for months, so I've been making Gemini tutorials for months, so I know exactly where it shines and where it falls apart, especially when it comes to text. OpenAI claims the new GPT images 2.0 fixes all that with better text rendering, improved detail, and even multilingual support. Big names. So I'm putting both tools head to head across three text-heavy scenarios, including one test most YouTubers completely ignore.