Nano Banana Pro is Here - New Features Unlocked!

Well, Google did it again. Hot off the heels of their Gemini 3 launch, they hit us with a new release that caught everyone off guard.

Surprise, mother. This time it's Nano Banana 2. Or as they prefer to call it over at Google, Nano Banana Pro. Now, I've had access to Nana Banana Deuce for about a week now, with Google demanding I keep it hush hush under severe threat of bodily harm. No, I'm just kidding. They actually asked me very politely not to talk about it until now. I got a business to run. I gotta kick asses sometimes to make it run right. But now that that embargo is lifted, I finally show you what's new with Nano Banana returns and show you exactly what you can do with this upgraded version of their most powerful AI image model. But before we jump into the demos, here's what it actually is. Today, November 20th, Google DeepMind officially launched Nanobanano Pro, their newest and most advanced AI image generation and editing model. And huge thanks again to Google DeepMind for that early access on this, because I've been poking at this non stop now. Nano Banana Pro is basically an evolution of their top tier image model. Inside the Gemini app, you'll see it labeled as Nano Banana Pro. It's built on top of Gemini 3 Pro image, which is why this thing feels so much stronger than the previous Nano Banana. Now, the idea is simple. They took everything people were already doing with AI images, mockups, posters, logos, product shots, explainer, graphics, style transfers, all that stuff and said, cool, let's make all of that way better, way cleaner, way more controllable and even easier to prompt. And not only that, but you're about to see it everywhere. Google's rolling this out across pretty much every platform they have. Here's where you can use it. First, for normal people like general consumers and students, it'll be inside the Gemini app. You should see Nano Banana Pro sitting right under that thinking toggle. If you're on the free plan, you'll get a limited number of pro generations before it bumps you back down to the previous regular nanobanana. It's also now powering image generation in searches AI mode, but only for Google AI Pro and Ultra subscribers in the us. For now, they're also rolling it into Notebook lm. If you're a workspace customer, you'll also see Nano Banana Pro inside the Gemini app, as well as baked straight into Google Slides. So if you build a lot of presentations, this is going to be a pretty nice upgrade. People who advertise with Google are also getting the Nano Banana Boost. Google is upgrading all of the image generation inside of Google Ads to Nano Banana Pro as well, so you can generate images for your ads directly in their platform. Developers and enterprises are getting the upgrade too. You can now use nanobanana Pro through the Gemini API, inside AI Studio and inside Google's brand new IDE called Anti Gravity. At the enterprise level, it'll show up in Gemini Enterprise and Vertex AI. So bigger organizations can run massive automated creative pipelines using this model as well. And if you're one of the people playing with flow, Google's AI filmmaking tool Nano Banana Pro is going to be available there also, but it's only coming to Ultra subscribers at first. So yeah, basically putting this thing everywhere. Google's also rolling out something brand new that I actually kind of love. Inside the Gemini app, you'll now have a feature where you can upload any image and ask, was this generated by Google AI? And because everything from their models have Synth ID watermarking embedded on them, it can tell you instantly that is part of that change. They're also going to remove the visible watermark on images for Google AI Ultra users, so you still get that invisible watermark for authenticity and traceability, but you don't have to live with a Gemini stamp on your image, which is going to be much cleaner. So here's what you're probably really here for. What is this new version actually capable of? Well, here's the simple version. This is the first time their image model feels like a serious all in one creative engine. Not just make me a pretty picture, but it can do clean, accurate, multilingual text rendering. It can design posters, diagrams, menus, infographics. You can blend up to six uploaded images. You can turn your doodles and annotations into real edits. You can create product mockups and brand guides. You can generate consistent style variants. You can do basically any aspect ratio to create images for any platform. And it actually respects the original composition now as well. You can apply style transfer from one image to another. You can reimagine group photos in different eras, genres or aesthetics. And there is so much more. I mean, now that it's available to the broader community, I'm sure all sorts of new use cases are going to start popping up from this one. But now that we know what it is and who it's for, let's dive in and actually play with the thing. My early access was provided through the Gemini app, so that's where all my testing will be. But you should be able to access it from Any of the places that I mentioned earlier. So let's go. AI videos have come a long way, but they still sometimes produce inconsistent results. So I want to show you a way to generate AI videos to get the best results. I'm going to use video because this method works really well on their platform and they just added the new VO 3.1 model. The first tip is to use reference images. This helps so much. You can find images on Google or have one of the other AI platforms generate them for you here. I just uploaded a reference image of Steve Jobs and another image of an iPhone 17 and the model just put them together. Trust me, this is going to look so much more realistic than if you just used a prompt alone. My next tip is to choose a model that's great at scene consistency, such as VO 3.1. So I wanted it to create a Mad Max style desert video. And you can see that the desert vehicle camera motion and the person driving the car look consistent between both scenes with no weird glitches between cuts. My last tip is to provide a start and end frame if you're doing transitional shots. This is similar to using reference images since it really helps the AI produce what you want instead of some random scene. To show this example, I uploaded a caterpillar and a Butterfly image and VO 3.1 filled in that transformation in between. You can try all of these tips right now for free@inVideoio. And if you're serious about video creation, they have a super advanced generative plan that starts at $100 a month. You can check it all out at the link in the description. Thank you so much to Invideo for sponsoring this portion of today's video. All right, so let's jump into actually testing this thing out again. I've had early access, so I've played with it, kind of put it through its motions and sort of know what it's good at and what it's not. So the first thing I want to test is advanced text rendering. Now they claim it can render text very accurately and in multiple different languages. So the first prompt I tested here was create three ultra clean modern logos that each spell the word wolf in different world languages with every letter made from hyper realistic materials like wolf fur, metal or glass. And this was our output. We got our metal, we got our woofer, and we've got our glass. I believe this is Russian and this is kanji, and I think this is Italian. I have no clue if these are correct or not. I'm pretty sure this one is Correct. Next up, I wanted to test the text accuracy, so I gave it the prompt. Design a bilingual cafe menu with English and Japanese in a minimalist style with perfectly spelled ingredient lists and prices and illustrated icons made from clay. This was the output here. And we can see we've got English and we've got Japanese. Now I don't know how to read Japanese. Maybe somebody in the comments could let us know if it actually did the Japanese translations properly. But sure enough, we have English, we have likely good Japanese maybe, and we've got images that look like they were generated out of clay for the actual items. Now let's look a little bit closer here and I'm going to read some of the text. So we have espresso, single shot of rich concentrated coffee. A lot of times when you have really small text in these images with these various image models, it doesn't actually read well. It's usually just gibberish. But on this, even the teeny tiny text below that describes the ingredients of this is all very coherent. Cafe latte espresso with steamed milk and light layer of foam. Sencha green tea. Traditional steamed Japanese green tea with a fresh grassy flavor. I mean, I don't see any issues with the English text at least. I mean, that's, that's pretty excellent. This is like an infographic maker right here. Like, this is pretty good. Now when I was talking to the people over at Google, one of the things they told me it was really good at was if you give it an image and then ask it to annotate the image with like facts about the image, it can do that pretty well. So I fed it this image of Petco park where the San Diego Padres play. And I gave it the prompt turn this picture into a detailed educational poster explaining the history, geology and fun facts about this exact location. And this was our output here. Petco park in downtown San Diego. History, geology and fun facts. I have no idea if the geology portion of this is correct. I would assume that's right. It did research to annotate this photo. If I actually show the thinking here on Nanobanana Pro, we could see it's analyzing the details and reviewed the accuracy and essentially did a bit of research. So this image, I'm guessing is correct. All the text looks good. San Diego Formation Sandstone stadium fill Bay Point formation Terrace deposits Artificial fill Rose Canyon fault zone. Petco park fun facts. It's the home of the San Diego Padres. Incorporates historic Western Metal Supply company building into design. That's definitely true. The Western Metal Building is Right here in the stadium, they've got the park at the park, which is a public green space. That's this whole green area here. This is actually an older photo. They've remodeled this park in the park. There's actually like a bridge here. Now that wasn't here in this photo, but that's correct. Seating capacity 42,000. Sounds roughly correct. The history of East Village pre 1850s title mud flats and Kumeyaay land 1880s to 1900s railroad and warehouses 1990s to 2000s redevelopment in ballpark construction. 2004 on Petco park opens. I mean, really good. I also did another test where I gave it the same image of Petco park and said, turn this photo into an informative visual explainer about this place. And it did a pretty good job. But where the arrows are pointing are incorrect. The text is all actually pretty good. But the San Diego Bay and Coronado Bridge aren't out here. This is looking like northeast of Petco park. And the bay is like. Would be down here. If you see where my mouse is moving, it would be like this side. If you've ever been to San Diego, you got Petco park here. You got the convention center running right along here. And then the bay is on the other side over here. So San Diego Bay and Coronado Bridge. It's literally the opposite direction. This one here looks good. Western Metal and Supply Co Building Historic landmark. It's pointing right here, but that's not where the building is. The building's on the opposite side over here where my mouse is. And then coaster and Amtrak railways. It's actually pointing at a parking lot right here, but you can actually see the train track going right along here. So. So the text is sort of over the train track, but the arrow's pointing to the wrong place. And then the gas lamp quarter. I'll give that one to him. This is the gas lamp quarter right here. This is sort of the main drag, this street here. So that one's pretty accurate, but a really cool way to make simple infographics out of images you already have. You can also doodle and modify photos directly inside of Gemini. So I uploaded an image here, and if I click on this image, you can see we've got some colored dots down here that actually let me doodle on this image. If you're wondering, you can see me right there. I was there. One little bit of feedback for Google is it would be nice if there was a way to clear drawings. So I just drew that arrow. I don't see a way to undo my arrow. Wait, wait, wait. There's an undo button up here. Nevermind, Google, you did good. So let me just draw like a big circle up here where it says made on YouTube. And then I'm going to pick a different color and let's sketch some weird stuff. Like, I don't know, I'll just make some random sketches over here and let me pick a different color and do this kind of thing and then go like this. I'm just making some random doodles and I'll click done and then I'll give it the prompt. Replace the circled area with something that fits naturally in the scene, but based on my doodle shapes, interpret the sketch creatively. Let's see what it does here. I mean, it's not really what I expected, but I also did just kind of give it some scribbles. Let me try something else. I'll give it the same image again. Click on our image and this time let's draw like a big giant in the background. I really suck at sketching with a mouse, as you can probably see. But let's go ahead and give it some eyes and some noses and some mouses and then, you know, some spiky green hair and you know, something like this with some hands. Yeah, just like that. Perfect. Another hand like that. And then I'll say giant person like this here. And then I'll point an arrow at our giant person. Click. Done. And then say add the giant person to the image. Interesting. So it added my giant stick figure here, but it also incorporated the stuff from the previous image, even though this previous image wasn't part of the prompt that I gave it. But anyway, the point is you can sketch directly on images and it will try to recreate that sketch. Make the giant person look realistic. Yes. Now. Now we are talking. This is what I was hoping for. Now we got a giant, realistic green haired person in the background of our image. Excellent. Let's get it to make a product mockup for us. So here's an image of a person wearing an Apple Vision Pro. Let's give it some horns like this. And let's make the eyes much more visible, like so. And I'll give it the prompt, use my doodles as instructions, and redesign this project with the features I pointed to. Make it look like a polished marketing image. I should probably point to some stuff, huh? So I'll go ahead and point to this stuff and. Perfect. Exactly what I was hoping for. Come on, you know you'd buy the Apple Vision Pro if it had those extra features. Right. So nanobanana Pro is supposedly pretty good at combining images. So here's a picture of a dog and a surfboard. And let's just say combine these images, let's see what it does. If I don't give it much instructions and we pretty much get what you'd expect to get. We got the dog and the surfboard in the same image. But here's where Nano Banana really sort of excels is. You could combine up to six different images all in the same prompt now. So I'm going to give it a picture of the dog, a picture of a surfboard, a picture of the San Diego skyline, a picture of a neon sign that says good vibes and a picture of a deserted road. And I'm going to get the prompt, blend these five photos and then I just sort of describe the photos into a single cinematic poster that feels like a sci fi road trip movie. Let's see what we get. I'm excited about this one. And we got a, a pretty epic poster. I mean it flat out took the back to the future DeLorean. But I mean we got our surfboard, we got our dog, we got our city skyline, we've got our road, we've got our like, it's all there. The long ride home, summer 2050. That's a long, long time to wait. All right, now I want to test six images. I'm going to drag and drop in six images all at once here of different, like clothing items. And I'll use the prompt, combine these six outfit photos into one unified capsule wardrobe style sheet showing mix and match looks. And I pick some random stuff and would you look at that. Here's our sheet. Let's take a peek here. So it took this hat, this blue shirt, this gold chain, these jean shorts, these cargo pants, and this skirt. And it created some mix and match looks that might be pretty interesting if you were seen wearing them in public. Now supposedly it's really good if you upload your own sketches and then turn them into something more realistic looking. But I'm not going to sketch anything. I'm too lazy for that. So I'm going to have it sketch something for me and then turn that sketch into something more realistic. Create a pencil sketch drawing of a futuristic tech gadget that will actually be helpful for people. So let's see what it designs that it thinks will actually be helpful for people. And here's what it generated. The agus. AGUS Health and Environment Monitor. So it's got a Nutrition scanner, stress and vital sensors, Air quality intake, nutrition scanner, a haptic alert system and stress at vital sensors. I mean, yeah, that's probably pretty helpful. Looks pretty futuristic. Now let's see if we can actually turn it into a realistic looking prototype. I'm gonna paste our image back in right here and we'll say transform this sketch into a photorealistic 3D render of a consumer tech gadget including materials, lighting and a realistic product hero shot. And just like that, it took this sketch that I totally drew myself and turned it into a pretty good looking realistic image. Need more concepts or mock ups? Let's try create a complete mini brand guide for a fictional beverage company named skysip. Logo, color palette, typography and sample packaging. And there we go. There's our logo for skysip. Here's the color palette. Even pick some fonts and we've got some bottle Sample packaging mockups. Pretty good. This new model is also much better at handling different aspect ratios. So let's do generate the same landscape scene in 11169916 and 3 2. Keeping composition consistent across ratios. So it generated this landscape image. This is our 1 1. We've got a mountainscape here at sunset with some trees and the reflection on the water and a little cabin here. And if we take a peek at our next image, We've got our 16:9 version of the same thing. It's just like a zoomed out version. Now We've got our 9 by 16 version that's a little bit more cropped in, but. But we can tell it's the same image. And then we've got our two by one version. All of them look like it's the same image, just different aspect ratios. All right, let's see how it handles converting an existing image's aspect ratio. So I've got this image here of me and Matt Berman. This is taken on an iPhone in 9:16. Let's see what happens when we tell it to change it to a 16 by 9. I'll give it the prompt. Resize this uploaded photo into a 16 by 9 wide poster while preserving all the key subjects with no distortion. And would you look at that? It look pretty good. Let's see. Curious about how it looked at the windows back here and sort of guessed what the rest of the building looked like. It assumed there was a third seat next to us, which they're clearly not. But again, not too bad. There's definitely some wonkiness. Like there definitely shouldn't be an extra seat there, but it did A decent job of widening that photo. Now Nano Banana Pro can also do style transfer so you can actually give it an image with a style on it and then another image and say take the style from that image and apply it to this image. So I'm going to upload this watercolor image here along with with this photo from Nvidia GTC that I have and we'll give it the prompt, apply the watercolor style of this painting to my uploaded photo while keeping all facial details recognizable. And dang, that's actually pretty impressive. Look at that. I like that a lot. It took the style from our watercolor painting and even some of the like color tones, like the same blues and oranges, and applied it to this image and damn, that looks good. All right, let's play with some color grading. The movie the Matrix has a very distinct like green color grade to it. I'm going to upload that same image of the San Diego skyline here and I'll say take the lighting and color grading from this film still and apply it to my city photo to make it look like a scene from the same movie. And not only did it recolor grade it, but it also changed the aspect ratio, which I didn't ask for, but that's an easy prompt to get it back to its regular aspect ratio. But it followed the color grade of the original image. Another fun use case is actually reimagining photos in different styles. So here's a photo that I have also from Nvidia gtc. I uploaded this photo with the prompt reimagining us as a retro 80s synth pop band complete with era accurate outfits, neon lighting and vintage album color typography. And well this is what we got from it. We got Matt Vid Pro on the keytar, me with my sick leather jacket, Igor from AI Advantage hanging out with his sports coat and Maria from Python simplified with that rocking 80s hair. And then one last test here. I want to see how it does with a big group photo and restyling that. So I gave it another image from the Creat collective with a whole bunch of YouTubers in it and said turn this group photo into a high fashion magazine cover set 20 years in the future with futuristic styling, makeup and wardrobe. And I mean it gave the futuristic background and futuristic text, but it didn't really change too much. I'm just chuckling here because I noticed that it gave Cleo abram here this like 80s retro futuristic jacket and this dude here's wearing like shiny armor. So I mean like some of the people it restyled, but it didn't get everybody. So I think it struggles a little bit when you give it like a giant group. Probably would have nailed it if there's only like four or five of us in there. But since it's such a big picture, it randomly decided who it was going to dress up. But there you have it. There's Nano Banana Pro, powered by Gemini, the brand new image model that just rolled out from Google. I'm pretty impressed. It's definitely, definitely the best editing image model I've seen that's on the market right now. Now there's probably some better image generation models, but as far as image editing goes, this one pretty much crushes everything right now. I'm sure it's only a matter of time before this also pops into Adobe Photoshop and Canva and all of these other platforms so people can just edit their images with text prompts instead of actually needing to know how to use things like Adobe Photoshop. Now, I didn't really cover all the other cool stuff that we've already been able to do with the original Nano Banana because. Because, well, we could still do all of that stuff. The original version was already super powerful. But if you want to go even deeper down the Nano Banana rabbit hole and find out all of the cool stuff you can do with it, check out this video that I made called 50 Plus Insane Ways to Use Nano Banana For Free. Because that really, really dives deep into all of the ways the original Nano Banana was able to be used. And today's video was more just covering the updated stuff that got better. But that's what I got for you today. That is the Nano Banana. This is hot off the presses. This video literally dropped at the same moment that Nano Banana released. Because I wanted to be one of the first people to show you all of the cool stuff that you can do with it. And if you want to stay looped in on all of the latest AI news, along with AI tools and AI tutorials, make sure you like this video and subscribe to this channel. I promise to keep you looped in on everything that's happening in the AI world so you never feel left behind. Thank you so much for hanging out with me and nerding out with me today. I really, really appreciate you and hopefully I'll see you in the next one. Bye bye. Thank you so much nerding out with me today. If you like videos like this, make sure to give it a thumbs up and subscribe to this channel. I'll make sure more videos like this show up in your YouTube feed. And if you haven't already, check out FutureTools IO, where I share all the coolest AI tools and all the latest AI news. And there's an awesome free newsletter. Thanks again. Really appreciate you. See you in the next one.