Nano Banana Pro Can Do WHAT? Watch These 25 Wild Examples

Google's new AI image model, Nano Banana Pro, is here and it's by far the best image model that I've ever used. And I have a list of over 25 examples that I wanted to show you in this video to show you how to get the most out of Nano Banana Pro. And in the description, I'll include the prompt list that I'm using for this video. Okay, so the best place to use Nano Banana Pro is inside of the Gemini app. So, if you go to gemini. google. com, you could use it right over here. and it is available to use for free, but it is extremely limited on how many images you could create with a free plan. With a paid plan, you get more usage. And I have something called the ultra plan, which also removes the watermark automatically. The first example I want to show you is actually changing the aspect ratio of an image, basically resizing it for different platforms. So, I have this image here and this is in 16 by9. I created it as a slide. And to use Nano Banana Pro, all you have to do is switch your model over here to the thinking model. So technically Nanobanana Pro is Gemini 3 Pro. That's what the image model inside of the new 3 Pro model is. So you want to choose this one. If it's on fast, it's going to use the regular Nano Banana, not the Pro version. And then under tools, you just have to go ahead and click this and then type out your prompt. And my prompt says resize this for 9 by6 aspect ratio. Okay. And here's the version we got in that aspect ratio. And you could see this didn't just crop the existing one. It recreated it completely from scratch here. And it kept all the colors, the logo, the font, everything is exactly like the other version that we had. So here's before and here is after. And you could always click on them and you could just download the full size right here from top of the page. They're also sharable over here too. So it will create a link you could share if you don't want to download them here. And this time I'll create a square one and I'll tell it the pixel dimensions 1080 by 1080. Let's send this out. And here's the square version again we created here from scratch. Now for the next demo I want to show you how to actually prompt it to create something with a specific size and aspect ratio too. So resizing is fantastic for something you've already created in Gemini or anywhere else. So for this prompt I'll say create a YouTube thumbnail in 16x9 aspect ratio. Announcing the release of Nano Banano Pro. Okay, perfect. This is the exact size that I need for a YouTube thumbnail. Now, no image model up to today was able to get aspect ratios right repeatedly. This is able to do it pretty much every single time. And now I could finally create YouTube thumbnails. Now, I have a different YouTube thumbnail example because it actually gives you much better results if you give it a reference photo and not design from scratch. But this is actually pretty good. Except the future of fruit tech is definitely something that should not be there. Let's see if you could actually remove that part. Okay, perfect. It kept everything identical and just remove the text. The future of fruit tech. Let me go ahead and download this one. Okay, the next example is creating head shot from an existing photo. So I have this photo right here. This is just a thumbnail image that I take and I cut myself out of the background, put myself on the thumbnail. I said, create a professional headsh shot using my image. put me in a suit with a blue tie. Okay, that's really good. And it followed the direction on the tie. I'm wearing a suit. I took off my hat. But now, let's have it put me in a more neutral background so I could actually use this maybe for a LinkedIn photo. Okay, that looks pretty good right there. Let me see if it could actually remove the background and give it to me as a transparent PNG. I haven't tried this one, but let's see if it works. So, it's kind of mimicking a transparency here, but this is not really transparent. If I download it and I open it with Photoshop and it's definitely not a transparent background. It's still a JPEG. So, I still will have to remove the background. Photoshop just has a one-click background remover. Now, this is a transparent background. So, I could actually move myself and put myself in a different background. So, still makes sense to use Photoshop for things like this. You could also generate backgrounds here directly in Photoshop. And I just used modern office blurry background and it gave me three different options right inside of Photoshop. So that step of using Photoshop for these type of things is still very relevant here cuz Gemini still can't create transparencies or PGs. Now I want to show you different demos on actually trying out different outfits. Then we'll look at merging different outfits that you already have like other images here. So I'm going to say put me in a tux at a wedding headto toe shot. Okay. So not bad. It also made me smile a lot bigger. It still kept my hat though. So that looks pretty good. Let's try something else. Let's say put me in a tennis outfit. Let me see if I could change my expression here. Not looking at camera. Change it to an action shot. Okay, not bad. It did put me in an action pose. And I asked it to make my hat white also in the prompt. But the physics of the game doesn't make sense. I'm actually playing the wrong direction, so they should not be in the background. So I could again fix that with a follow-up prompt. And this time I said, "Change my direction so there are no players behind me and put me on a

clay court sunset behind me. " and it followed every piece of that prompt here. I mean, it put me in a clay court in the middle of a desert. But again, that's why it makes a lot of sense to be very specific with your prompt because it will just make assumptions every single time. But if you define exactly what you want, you could get it to do exactly what you're asking most of the time. Okay. Now, I want to show you how to actually merge several different images into one image. The example I'm going to use here is I'm going to upload a picture of myself and I just took a couple of screenshots just from the Gap website here to see if it could put me in this outfit. Okay, that's not bad. It's definitely that exact sweater here that I took a screenshot. Yeah, that looks great. The pants are right. It followed the prompt. Also, I said, "Put me in a white background in a photo studio. " So, it followed that. And I have a goofy expression on my face because again, I did not define that in my prompt. It just kind of defaults to whatever you were doing before. I guess I wasn't even smiling that big in my original image, but it keeps making me smile like this. Now, I have a few more examples of merging many, many different images and getting one shot out of it. But let me just switch over a little bit to show you a couple examples that are workrelated. So, I create lot of infographics because we post them online across our different social platforms and lot of different use cases for infographics. So, let's see if it could create an infographic for us. So, here's a text prompt for an infographic that explains the flow of generative AI. Basically, how generative AI works. And then I give it some details on what I'm looking for. Include reasoning models, include icons of each stage, arrows moving from step to step. Let's see what we get here. Okay, that looks pretty good. And it actually did a good job putting this in the right order on how these type of things work. Now, as a follow-up prompt, I'm going to type in another prompt for an infographic, but I'm going to ask it to create it in the style above. And that's all I'm going to add in the style above. Because if I start a new chat and I put this new prompt here, it's not going to follow the style I have above. And it followed exactly that type of style here. But this is something completely different. This is the evolution of machine learning starting from the 1950s all the way to near present time here. And again, you could see how good these models are getting with text. Just about a year ago, they could not do text. No model could do text. Chat GPT's image model couldn't do text. Midjourney couldn't do text. Now, it's almost perfect in every single prompt as far as you define what you want in the text part of the prompt. Now, the next example is turning raw data into visual graphics. So here is just some madeup data about a company revenue here. Quarterly revenue, expenses and so on, right? But let me ask it here to turn this into a visual graphics. And here's the image is created for us. So it added the company name that I gave it. This was 2025 quarterly revenue and expenses. And it was the numbers on every quarter and I checked every single part of this is correct. The numbers are correct. The way broke it down is correct. total revenue is correct. So, how great is this then to take this inside of an app to create a presentation for example? Now, I created a lot of these type of graphics to use in presentations both internally and sometimes externally where I'm creating courses and things like that and I use Gamma for that. Now, Gamma is an awesome presentation tool. I wanted to show you an example here, but it's also another place where you can use Nano Banana plus over 20 of the latest AI models, image AI models, and I partnered with them for this video to show you exactly how it works because I think it's a very practical example here, combining the power of image creation with a really nice AI powered presentation tool. And Gamma is a free-touse AI design partner for creating really effortless presentations. You could also use it to create websites, social media post, and a lot more. And it does not require any type of design skills or coding. And that's why it makes it one of my favorite apps of all time in the world of AI. And I'm definitely not the only one. Over 50 million people are using Gamma and it's the most popular AI presentation app in the world. And on the Gamma website, you could create a presentation multiple different ways. One of my favorite options is pasting text, but you could generate from a single text prompt or you could import from any existing presentation or even web pages. I'm going to use this option right now. And I have just some raw data here, but you could put in as much detail as you want. And Gamma puts it in this card by format for you automatically. I did not have to define which part goes into which card. So, it makes it super simple. And then you could generate based on how much information you want on the card. Now, here's my favorite part with these. You could actually generate images here on the fly. Gamma will take care of it in the background for you. And you could choose the image model you want to use.

If you haven't used these before, Flux is great. Ediogram is great. And if you have the premium plan, you get access to other models. And if you have the Ultra plan inside of Gamma, you also get Gemini 3 Pro, which is Nano Banana Pro that I'm showing you here. So, you get access to that as well. So you could decide which model you want right up front and it will generate all the images and in a few seconds this is in real time right now. Gamma is creating this entire presentation right in front of you right you could see it's generating every single slide and it's also generating the images. So all the images that you see are generated here in this case with Gemini 2. 5. But I want to show you a really fantastic thing. I think it's pretty much almost done already and I have a presentation that I could download. But they also have introduced something called this AI agent here. This gamma agent will make changes for you from a simple text prompt. I'm going to say change every image to actually use the data from my presentation. So it will change every single image in your entire slideshow. And just like that, every single image is turned into an infographic here. So it fits my overall presentation a lot better. Now I could directly present it from here. They have that option right inside of Gamma or you could also share it here. And they let you export this as just about any type of file you want. So you could do a PDF, even a PowerPoint if you want. And you could also publish this to a website. So if you're still using PowerPoint in 2025, in 2026, I think it's really worth trying out Gamma. I'll put a link in the description so you could try it out for free. This next one is a fantastic tool for marketing and any type of e-commerce. I'm going to add a logo to any existing product. Let me show you how it works. Okay, so I created this image here, but I want to actually change the logo. So, it created a logo here, but I want to add our Futuredia logo. So, I'm going to say replace the Lion logo with our logo. Okay, here's the logo right on the can. But now that the logo is there, we could put this product in bunch of different scenes. Here it is inside of a gym setting. Now, this time I put it on a track and I added a person here. Here's a version of it with a gamer. And pretty much every time I just say put it here and then I describe the setting and it just puts that same drink in the same conversation in different scenes. Okay, so it decided to use Michael Jordan for this one. So maybe we want to be more specific and use a generic basketball player because I don't think we just use Michael Jordan here to promote this madeup drink. But as you could see in the world of e-commerce, this has made creating posts for social media or ads a million times easier. For the next example, let's go ahead and see if it could put us on the cover of things. We'll start with a Time magazine cover here. And I didn't give it any prompt, but I'm going to follow up with a prompt to add some text to show you how good it is with text here, too. Okay, that's not bad. The text is not great because it's cutting off the word feature here. So, we'll go ahead and revise this. And another thing I wanted to point out is if you have a conversation where I just created this headsh shot and I created that PNG transparency and all those things, I shouldn't continue that conversation. This is where Gemini, even with this version of Nano Banana Pro, makes many mistakes like right now it still kept things related to my last prompt, right? So I don't want that. So I'll start with a brand new prompt and a brand new chat. Okay, here's the cover of Time magazine. Next leap in AI is here. This text is gibberish. So prompting actually still makes a big difference. You want to spell out as much of it as you can to get what you want. It also used my image exactly from the thumbnail. It has not color corrected me to match. So it's still going to require a few other follow-up prompts. Maybe put me in a different pose or have me actually maybe wear a different outfit and things like that. So, so far it's great at following the prompt if you really spell out the prompt. But if you keep it open-ended, it will make assumptions and a lot of times it's not really what you want. So, I recommend make your prompts a little bit more defined than just a single sentence. So, for this next one, I'll define it a little bit more. Put me on a cover of a movie poster. The movie is called The Leap with two other co-stars. Put me in a cyberpunk outfit. Make the poster futuristic. bold release date that says November 15, 2026. You got this part right. It made up against some text that doesn't really make sense. So, every time you define the text, it usually gets it right with the correct spelling. If you don't define the text, it just puts gibberish there. So, I recommend taking any prompt you have and just using Gemini to ask it to flush out the prompt. So, I'll give you an example here. So, I just said, "Redo my prompt below and give me a new prompt in text format. " so it doesn't try to recreate the image again. Add a lot of details including every part of the text that appears on screen. And make the photo realistic. That was a little cartoony for me. And make it poster size. So it just defines the aspect ratio here. And I included my prompt. So I'll let this go to work. And here is a new prompt. This optimize text prompt. So it's a whole lot more text. It still followed

the things I wanted like the title is there, the release date is there, but it also added the other text I want. So I could go ahead and edit those here in my prompt to make them very specific. And it gave me all the placeholders here. And it also defined a style. Cinematic lighting, for example, high contrast. Okay, this one followed the prompt a lot closer. It also did not make up any text because we told it very specifically what text to put here. So it did follow that. I still think it's not quite ultra realistic, but it's not bad. And I may want to change some of these things in the background. Okay, let's look at a few demos here that requires more advanced photo editing. So, this example here is merging two different photos. So, I have this one here of a beach and of stars, Milky Way, galaxy here. So, I want to merge the two where the beach gets this type of a sky. And here's the result. And all I did was upload those two photos and I said merge the two. This usually takes some skills with advanced photo tools like Photoshop to be able to blend these two so seamlessly. Now, let's see if you could actually correct colors. So, this one I messed up the colors on purpose in Photoshop. I also took the contrast and brightness way down here. And I'm going to ask Gemini to fix this photo. I'm not going to say fix the brightness. color. I'm just going to say fix the photo. See what it comes up with. Oh, interesting. I got my first warning here. Normally I can help with this type of thing, but I don't seem to have access to that content. That's strange. Let me try again with a new chat. Okay, so it worked that time. Let me look at this one. Yep, that looks really, really good. It fixed the contrast. It fixed the brightness. I mean, compared to where we started, and it's exactly that person. So now, can it fix things that are out of focus? So images that are blurry like this. Now, I did this one on purpose, and it's pretty blurry here. You can't see a lot of the details. Let me see what it comes up with. Okay, that is shockingly good here. I mean, look at them side by side. It did a really good job. I want to do one more. I want to really increase the blurriness and see if it could get that. And I started a new chat, so it doesn't have any of the reference from the old chat. Okay, that is just wild. And I worked as a cinematographer for a long time. If you missed focus to that level, it was never fixable with any tool. Okay. Now, can it remove objects from a photo? I want to see if it could remove just this one car. So, I'm gonna ask it to remove the second car in this image. Okay, it removed the second car. Let me just look at the original image here. It looks like it also altered the car behind it. So, it kind of removed two cars, but it followed it pretty well. Let me see if it could replace it with a red bus. Now, my prompt said put a red box there. So, I specifically meant here, but it put it on the other side. Let me see if I could change that and put it on the left. Okay, it got it right this time, except the bus is going the wrong direction. I think this is the back of the bus. It's the exact same one that was over here going the other direction. Right. So, again, with precise prompting, you can get it to follow your lead, but you can leave things open-ended or you'll have to fix it in follow-up prompts. Now, let's see if you could colorize an image that was taken in black and white here. So, let's ask it to colorize this. Okay, that's pretty good. Let me just put them side by side here so we could compare. Yeah, it looks really good. It maintain all the details here. Now, let's see if it could restore an old image. Here's an image from the 1920s. I'm going to ask it to restore it. I mean, yeah, that's pretty good. Still the same people, but it restored the image. Let's see if we could modernize it and colorize it, too. Pretty good here. Messed up the face a little bit there, but overall, I think it's passable. Okay, now let's actually create some content for social media. Let's start with an Instagram post. So, I have this one. Design a square visual guide called AI tools entrepreneurs should try. Okay, it's pretty good, but my prompt actually had a little bit of a flaw. I asked it for AI tools, but then I asked it to give me categories. So, it followed my prompt exactly, but it's not quite right because I need to add specific AI tools, which I could fix in a follow-up prompt, but it followed the size for a Instagram post here. So, I think it's pretty good. Let me see if it could change the entire style of it, though. I'll upload an image that we created before right here. And I'm going to say recreated using this style. Okay, it's not bad. It followed it to some extent here. But a better way to do this is not with a follow-up chat. Actually, start your chat with a reference photo and ask for your prompt to include the pieces that you want included. So, let's do that for the next example here. Let's try to create a Facebook ad. And what I'm going to do here is instead of taking a screenshot, I'll just go to my website here and I'll copy the URL. Let's see if it could create a Facebook ad from a link. in square format to promote my company. And I just pasted the link here. It did

create an image. Level up your future with AI skills. So I figured out what we do from the link, but definitely not a great ad. Again, this is where a reference photo goes a long way. So I'll give it this one. This is an actual ad we run on Facebook here. So, I'm going to see what it comes up with in this style. Okay. So, you could see it could recreate that exact same style, but it's not done a good job here coming up with a headline or a good CTA. It's changed the cards here. So, this is where it makes sense to use chat GPT or Gemini. Flush out exactly what you're looking for and then tell this Gemini model to exactly create that because if you try to combine it to do multiple things, it just doesn't do a good job. It's not using Gemini in the best way to come up with different headlines and it's not using Nano Banana the best way to follow that instruction. So that's why a couple of different tools here do make sense for Facebook ads. Now let's try one for LinkedIn. Design a corporate style visual called AI integration roadmap for business. And let's see if you can make an infographic that is more appropriate for LinkedIn. Okay, that's not bad. But it's very cluttered here. I think it's too much text. I'm going to ask it to kind of clean this up and use a lot less text. And yeah, that looks a lot nicer. And again, with a visual reference, I could get it to copy another existing visual reference to make something that matches our brand and our style and our font. So, I always recommend start with something you already have, use it as a visual reference, and then give it a prompt to create something for you. But this looks pretty good. Now, let me show you how to create YouTube thumbnails in the most effective way. Because the example I showed you where you created just from a text prompt was not great. The best way to do it is if you give it a reference photo. So this was the reference photo. And I'm just going to say recreate this thumbnail but replace the text with Nano Banana Pro. Okay. How about that? How good is this? It looks incredible already. Now let's do another revision. And I said have me hold a giant banana. Kept everything else the same. Changed my hands here. And now I'm holding a banana. This could be a winner here. You could also use it to create memes. So create a meme about a stressed out business owner buried in bunch of paperwork and so on. Let's see what it comes up with. Oh, that's pretty good. I actually think this makes for a better Facebook ad than my Facebook ad example I was trying to put together. Now, can we actually make a website landing page? So, I flushed out my prompt here with a little bit of help from Gemini. creates a clear and engaging landing page explaining what Skill Leap offers and it created this landing page for us. It's okay. It's definitely not anything I would use here. Vibe coding apps actually do a far better job and you could actually use them right inside of Gemini. So, instead of an image model, you could use another tool here called Canvas and then give it that same prompt here and it's going to make the actual website for you or give you the code at least. Okay, I got this mockup using canvas this time, not the image tool. So I just wanted to show you an example where the image tool is not obviously always the best choice. In this case, Canvas design us something much better. Now I also made this resource for our skill leap members, but I made it completely free right now. So you could jump into our website and I basically put all the different examples and the prompts that I use for each example here. So you could see kind of what the example does and then the prompt that you could copy and paste. A lot of them are templates, so you could add your own placeholder there. But pretty much everything we covered in this video is here, and I'm adding a whole lot more to it by the time I publish the video. And it's all searchable. So if you want to look up something like logo creation, you just search for it there, and it links to the top AI image models like Nano Banana directly from here, too. So I'll link that below, and it's free to use. And I wanted to show you one last resource here. We have an entire course on Nano Banana on Skillap, our learning platform. and Kevin and Andrew teach that course and it comes with prompt books, a lot of different practical examples and we're adding more examples from Nano Banana Pro that is going to be added to this course very shortly. And we also have a learning path. So this learning path creative guide for Gen AI has some of our best courses when it comes to creating images and those type of things using the top models like midjourney, the chat GPT image generation. We're adding nano banana here. There's other courses available as well in linear order. So, you get access to all our courses and learning paths with our subscription. And we have a free 7-day trial to that unlocks everything. I'll put a link in the description to that if you want to check it out. And I recently covered everything that came out with Gemini 3 Pro.