ChatGPT 5.2 vs. Gemini 3 Pro (Head To Head Test)

So, OpenAI just dropped GPT 5. 2, and this is the moment we've all been waiting for. Gemini 3 Pro has been sitting on the throne for weeks now, crushing every single benchmark and converting so many users from Chat GBT over to Gemini. But this is OpenAI finally firing back. So, today I'm putting both models through a head-to-head showdown across multiple realworld tests to show you which one actually deserves to be your daily driver. Let's get into it. This video is sponsored by IT 10X. Now, OpenAI is

claiming that GPT 5. 2 is the smartest generally available model in the world and in particular is good at real world knowledge world tasks and industry experts prefer the model's output to the output of other industries uh industry models aka Gemini 3 Pro. If we look over at the benchmarks, GBT 5. 2 2 is now

outperforming the competition in almost every single category. This means it should hallucinate less in logical tasks. It should have a much stronger ability to handle new problems that it hasn't seen before. And finally, it does appear to have a significant lead over Gemini when it comes to general knowledge work. But personally, I never put too much stock into benchmarks. Let's see how it actually performs in the real world. So, for this first test

I really want to test how fast the new model is and also whether it hallucinates and how good it just is at logic and understanding. So, I'm going to give both Gemini and Chat GBT this trick question that basically asks it to do a calculation based on some mammals on Mars. Now, obviously, there are no mammals on Mars, so it's not going to be able to do this. Both of them gave an answer, I would say, in about the same amount of time. It was there. Both of them were very fast. You can see right here though, Chat GBD's answer is much shorter, but both of them are correct. They both correctly identify that there are no native Martian mammals. Now, looking at Chat GPD's answer, I'm pretty surprised because 5. 1, the model before this one was very verbose. If you asked it any sort of question, it always gave like a really long answer that I personally hated. I just wanted it to kind of get to the point. And this has done that. It got to the point very quickly without really overexlaining things. Gemini here also gave a great answer, but there's a lot more information in here, right? So, let's see if 5. 2 continues to give these more concise answers because personally, that would be a huge step up for me. So, next

I want to test out its multimodal abilities by uploading an image and having it analyze that image. So, this is a photo I took recently at a Warby Parker store. I was shopping for some new glasses. This was a pair that I particularly liked. So, I took a photo of it so I would remember it. And I'm just going to ask both of them to look at the photo, tell me what's in it, what I would probably what anyone would probably want to know about the photo, where it was probably taken, and why it was probably taken. We can see that chat GBT is moving faster than Gemini. It's already creating an answer for me. We're still waiting on Gemini over here. So this says a retail eyewear displaying the waqen 175 $175 wide. So was properly able to identify text in the image. Now it didn't exactly say that it came from a Warby Parker, but it said it's likely a mid to premium brand like Warby Parker or Lens Crafter and it's a brickandmortar retail environment. And here we go. One of the reasons it might have been taken was so that you could remember the model name and price or maybe you were comparing styles. then gives me some extra useful details like it says it's $175 which are likely the frames only. The lenses cost extra which is not true for Warby Parker and is someone with a larger head size which is true. So not bad. It would have been great if it was able to identify Warby Parker. So over here uh Gemini said this photo displays three shells of optical glasses. These are the waqen 175 wide so is able to pull text from it. And this one was a lot more certain. Gemini says this photo was almost certainly taken inside a Warby Parker retail store. And it explains why because it was actually able to pull the information from the Warby Parker website. And it was probably taken to remember the style or for a fit comparison or maybe to ask your partner or friend which of these styles do you like the best. So of these two answers, I think Gemini's is definitely better. However, I will say that both of them did an equally good job of actually identifying what was in the photograph. I don't think there's really too much of a difference there. It was just Gemini was able to do a lot more in identifying exactly where it was taken and doing that extra work. Now, let's test 5. 2's

ability to create images. And this is one I'm very interested in because I've been absolutely loving Google's Nano Banana, their image generator. And I think OpenAI has a huge gap to overcome just how good Nano Banana is. So, I'm gonna ask both of them the same prompt, asking it to create this futuristic street market with a robot selling a bunch of glowing blue apples, a bunch of people milling around, a bunch of robots signs in the background as well to just make this as complicated as possible. And here we go. Gemini is already finished and Chat GBT is not. So, let's go ahead and take a look at this one. Uh, I really like this. It's followed the prompt exactly. We have all the text really well written out. Neotokyo noodles. That one it came up with on its own. Quantum repair came up with that one on its own. This little cat over here. All these people and robots raining. This is awesome. I mean, it's it looks like New York City, which is what I asked for. One thing that Nano Banana has always been good at is putting text on images. So, I'm really not surprised with how good this came out. Meanwhile, Chat GBT is still working. And finally, Chat GBT is finished. This took a couple of minutes. Gemini is way, way faster still than Chat GBT, but it did a good job following the prompt. We have the robot, fresh parts, there's the cat, the glowing blue apples, lots of signs in the background, though they're a bit repetitive. A lot of them say hotel, restaurant. So, you know, overall, I think Gemini's looks better. Uh, I just think it it's a lot nicer. There's a lot more going on. You can see a lot more. I think it's better. But, let's go ahead and actually test some other image generations with this as well. Let's now

test their ability to edit photos. So, for both of them, I've given them two images, one of a female model and a pair of sunglasses. And I'm going to tell both of them to put the sunglasses on the woman. Now, ChachiBT is still analyzing the images. Meanwhile, Gemini is already finished, but let's see how good it looks. I mean, that is spot on. Oh, no. It messed up. Look at this. Look at the ear. That's pretty surprising. You know, I've been using that this for a while, and I haven't seen mistakes this obvious inside of Nanoo in a while. So, not perfect here. These sunglasses are not quite going into her hair correctly, but this maybe is a little bit challenging. Let's see how well GBT 5. 2 handles this. Otherwise though, this is really good the way it's sitting on her face. And these are the same sunglasses as well. I don't know what's going on, guys, but I've tried this three times and can't get it to work. So, I guess Gemini is the winner of this

one. So, this time we're going to do a multi-step edit. So, I'm going to tell it a prompt that I use all the time to create my thumbnails. I use this in Gemini all the time. And so what it's basically asking for is a viral YouTube thumbnail. So it has to understand the aspect ratio of that. what goes into a viral thumbnail. And I gave a specific instructions on what I'm looking for. So Gemini produced a result for me very quickly and overall I am very happy with this. This is exactly what I asked for. These are the exact words. This is the exact image I wanted. On the right there's my face. case. I told it to make me look excited and pointing at the laptop screen, which it did. I told it I wanted a YouTube style background, which it did. So, this followed my prompt exactly, and it's nearly perfect. It got all those nice details in there. Meanwhile, GBT 5. 2 is still working, which at this point is to be expected. Oh boy. All right. That's rough. That's really rough, guys. That's bad. That's really bad. Yeah, I don't even know what to say. What do you guys think? Let me know in the comments section down below. But I think when it comes to image generation, Nano Banana Gemini still blows Chat GBT out of the water. No competition. Okay, so let's

move away from the images and the image generation. This time I'm going to do some creative writing or some writing with it. This is something I do every single day. I actually have it help me write out the scripts for my YouTube videos. I don't script out the entire videos, but just like the intros and outros, which is exactly what I'm going to ask both of them to do right now. I want them to craft for me three hooks for this very video. Now, both of them again very fast. And what I'm seeing again in chat GBT is just how much more concise it was than in the past. Okay, so I've read through both of the responses and I got to say GBT 5. 2 is much better. the ones I got from Gemini. I've never really used Gemini for any sort of writing and I just kind of feel like it's a bit traumatic at times. So, this one for instance, it says Gemini 3 Pro reigns supreme, but its time at the top might be over. This isn't just an update, it's a declaration of war. Where this one says, stop scrolling before you pay for your next AI subscription. You need to see this. Right? Just a little bit too much hype, dramatic. And GPT 5. 2, 2 I think actually gave good responses here. Like this one at the top says, "Not long ago, Gemini 3 Pro basically dethroned Chat GBT. A lot of people, including me, started switching, but OpenAI just dropped ChatGBT 5. 2 and this is their direct response. " The real question is, did they actually take the crown back? And that I think sounds amazing. That is excellent. That is a hook that I would actually use. And the other two hooks that I came up with are also excellent ones. I also think the way that it organized the information is just easier to see over here on Gemini. It gave like a title. It gave a little bit of an intro and then the prompt right here. I guess it's nice that the prompt is sort of is off to the side like that, but I think it's just like too much information here. This is a lot more concise. Uh for me, Chat GBT wins this one hands down, which means it's

actually getting harder to choose now between Gemini and Chat GBT. But luckily, because of today's sponsor, you don't have to choose. This video is sponsored by i10X. And honestly, the timing couldn't be more perfect because what we're seeing right now is the exact problem that i10X solves. Look, if you want to use GPT 5. 2, that's going to cost you $20 a month. If you want to use Gemini 3 Pro, that's going to be another $20 a month. Throw in Perplexity Pro, that's Claude for writing. Suddenly, you're spending well over $100 a month just to have access to the best AI models. But iTX gives you all the best models in one platform. And it starts at just $8 a month when you do the annual billing. You can see right here that inside of i10X, when you come over to the chat, you can actually choose which model you want to chat with. And they have all the top models in here. Anytime a new model drops, as soon as the API is available, they add it into i10 X. So for 5. 2, that means this will be coming very, very soon, as soon as it is available. So no more switching between different tabs, no more managing multiple subscriptions, no more deciding which model you want to use because you can only afford one of them. All of them are available here inside of i10X. And one of my favorite things about iTX is this feature right here called chat arena. And this allows you to compare two different models against each other in real time, exactly what I'm showing you in this video. So I can take one of OpenAI's models and I can compare it against let's say one of Google's models. And then you can type one prompt in here, click on send, and then both of them are going to get to work at the same time. You can see which one is faster, which one follows instructions better, and which one produces the better results. and then you'll know which AI model is the best for the job that you just described and you can continue to use that in the future inside of AI chat. On top of that, you get things like AI image generation. They have all the top models in here like Google's Nano Banana as well as a bunch of other ones. They also have deep research which is Perplexity Sonar Pro. You can chat with PDFs and documents. You can create videos with this. And we even have an agent builder in here as well. This thing is really powerful. This allows you to create these pre-built AI assistants for specific tasks. So, if you're tired of paying for multiple AI subscriptions or you've been sitting on the sidelines trying to figure out which model to actually commit to, I 10X might be the answer for you. It gives you access to all the top models in one place for one low monthly fee. So, if you want to try it out, click the link in the description down below. And remember to use code PJL15 to save an additional 15% off. All right, let's get back to the tests. So

now I'm going to ask both of them to create for me a landing page for my new business, which is an AI course. So, this is really test out its coding abilities. Now, I'm not a coder myself, so just take all this with a grain of salt. Right away we see that chat GBT is moving much faster. Now we have Gemini sort of catching up doing all of its coding. So both of them are definitely coding as to be expected. ChachiBT finishes first. So let's go ahead and preview what I created. Guys, this kind of sucks, not going to lie. I mean this does not look like a landing page. This is like ton of text up here at the top. Yeah. Like you can't even read this. Like it should have known that. I mean this is totally just like unworkable. Wow. Meanwhile, Gemini, it finished up over here and this is what a landing page is supposed to look like. Nice big text up here at the top. And then we have the buttons right here. Scrolling down, you can actually see everything. Now, right here, I told it to only make three of these. That was inside of the prompt. It would have looked better with four and it would have been nice if it was able to, you know, style this in a way so that three looked nice. Maybe if I allowed it more flexibility, it would have added the fourth one which would have looked better. Simple pricing down here. Uh, this is a lot better. better than GPT 5. 2 made. So, I was

lot better than GPT 5. 2 made. So, I was honestly so disappointed in those results I decided to give GBT 5. 2 another chance. So, I asked both of them to make for me a very simple task app. And again, Gemini just blew GPT 5. 2 out of the water. This is so much better than the one that GPT 5. 2 has made just in terms of the way it looks and the UI. Let's see if both of them work though. Out trash. There we go. I check one off. We got some confetti. So, this one I could do take out trash. And yeah, it works. Yeah. So, both of them work the way they should, but the Gemini one just looks better. It's a lot cleaner. You know, this is what I would want. Not this one on the right. Next, I want to

test out just how good each of these models are at just chatting with it. Like if you're just bored or just needed someone to talk to, like how each of these react to this. And Gemini, I have always found just feels a little bit sterile to me. Every time I chat with it, it kind of feels very corporate and like it's trying to help me. like it understands it is an AI assistant who's trying to be helpful more than anything. Meanwhile, chatb 5. 2, I'm happy to say, really feels like you're having a conversation with someone. It's a very natural back and forth. Now, just like it always does, it always ends everything with another question, which kind of feels uh a bit tedious after time just getting all these questions at the end of every single response that it gives you, but it's always done that. And I'll just say that the answers it gives in general just yeah, they're great. I think for general chatting, GBT 5. 2 is definitely the winner here. For

this next one, let's see how well they both do at strategizing. I'm asking for each of them to help me with a new online course. I need to know which videos to make, what topics I'm going to talk about, the emails I'm going to send out. Again, you see chat GPT just cruising along so much faster than Gemini already spitting out an answer for me. And here comes Gemini. So, let me read both these over real quick. So, comparing the two of these, what I'll say is that I like the way Gemini organized the information better. It's just a little bit easier to understand how everything works together because it really broke it down by these phases and then within each of these, it has these tables. So, it's just really easy to see this information organized. ChatG didn't do as good of a job as that. However, it's still understandable. You can still see how it all fits together. for this entire timeline they put together for this product launch. And I'll say that chatb did a better job with the sales copy. Just like we had with the YouTube script we put together, chatbt I think just sounds more natural. The titles that it has suggested for the videos are better in my opinion. Again, the Gemini ones just kind of feel a bit overhyped or a bit sensationalized where I like the ones from Chat GBT better. So, I would give the win here to ChatJBT, although I really wish it had organized the information as cleanly as Gemini did. And finally, let's test both of

their abilities to look at PDFs that we upload into it. I asked both of them to find three vegetarian recipes from the PDF that I uploaded. GBT 5. 2 is much faster. It has found for me three, which is a broccoli kiche, a leak tart, and tomato and fontina cheese tart. And of course, it references the PDF that we just uploaded. And it says if you want, you can pull the dessert only, vegetarian recipes. I think most desserts are vegetarian at least, even vegan friendly ones or whatever else. So here we have Gemini finally coming up with all three answers. And this one actually not only gave us the name of the recipe, but also actually gave us the recipe itself. We see three of them, including that broccoli kiche. Uh, it's a little bit of a messy answer because for the first one, it actually gave a full breakdown of the recipe, and for the other two, it didn't, which is a little bit weird. So, yeah, Chhat GBT wins this one. It's much faster. And of course, I could follow up and ask it what the recipes are. Now, obviously, this is just my initial reaction to 5. 2. It literally just came out 3 hours ago, and I do want to do a lot more testing with it. However, my initial analysis is that I'm sort of underwhelmed by it. I was expecting a much larger leap the way it was being hyped up by OpenAI, and this doesn't feel like that to me. I will say that any sort of text output that it produces is very good and I think it's better than Gemini and it may end up being my favorite model for things like doing creative writing or helping me with emails, coming up with titles for my YouTube videos, anything like that. But it still lags way behind when it comes to image generation. And I just have to say, even though I'm not a coder, so again, take this with a grain of salt, the coding abilities seem to be lagging behind Gemini as well. But again, that's just coming from a casual coder who's just doing some vibe coding like you saw here today. So, I definitely want to test this a lot more over the next few days and weeks. So, definitely subscribe to the channel to hear my future thoughts on it once I've actually been able to test this and really use this in my workflows a lot more. And also, please let me know in the comments section down below if you think I'm missing something or if you've tried 5. 2 what your thoughts are on it because I would love to hear from you guys. That's where I learn a lot as well. And again, maybe there's something I'm missing here that I didn't realize that 5. 2 could do or some of its capabilities that I haven't really tested here today. And again, if you are just sick of this back and forth battle between the two of them and just want one place where you can access both Gemini and GBT 5. 2 and all the best models out there, make sure to check out i10X, the sponsor of today's video, which I'll have linked up in the description down below. Otherwise, guys, thanks so much for watching and I'll see you in the next video. Bye for now.