Google's NEW Free AI Image Update is INSANE 🤯

Google Gemini have just unleashed the Kraken today with their new update, Google Gemini 2. 5 images, and it's gone bananas, aka Nano Banana. So, if you're not familiar, basically there was a release earlier this week where Google were talking about there was this stealth release of Nano Banana, it's called. It was a really good image editing model and it's been revealed I think as we all suspected that this is actually Gemini 2. 5 images. Now there's

a few ways you can use this right if you want to get started you can go straight to Gemini like and you can just select images you click on the tools and then create images from here and then you can start generating this right. So for

example, I can say mafia cat in a suit eating pasta but keeping the same character consistent, right? So for example, we got a cat eating spaghetti right there. I mean the time of its life, right? And then I've said make the hat a bowler hat plus in the previous prompt I said different cities every night, right? And so now it generated a mafia cat in Bangkok enjoying some pasta sporting a bowler hat. And so it's really easy to generate these images. is

far faster than using something like chat GPT. We can actually compare the outpas here. Here we have Da Vinci painting the Mona Lisa. This is Julius Caesar reading an iPad in the Roman Senate. And it's really easy to generate these. Here's another example. So this is a cat explaining generative engine optimization on a whiteboard like a professor. And then we've got the cat having a great time just presenting that. So if you want to generate more cats which the internet knows it needs then you can do that using Gemini images. You can also say for example Muay Thai fight and Bangkok with the fighters yetis and suits and then you could say okay change the yetis to Bigfoot plus make the image feel more like a photo not a cartoon. Right? Something like that. And it's really fast when you're using this. Plus, this is completely free to use, right? So, the brand new update is completely free for anyone to start using and you can use it straight away. So, now we've got the image. It does seem to struggle with generating like photo realistic stuff, but you can see here it switched people to Bigfoot in suits instead. And then if you want to, you can download this.

And you can also use this for generating videos, right? So, for example, if we start a new chat here, we can actually add the photo we just downloaded. And then we can say, okay, turn this into a video and we've taken the photo generated with 2. 5 images and then we're going to generate a video with Vo3. So, you've got like this allin-one media production process using Gemini images and 3 Gemini video generator inside Gemini, right? Nice and easy. Now also

what you can do here is you can go over to open router and you can get the API for it's already out. So if you go to open router models and then scroll down you can see there are two different options right here. So you've got Gemini 2. 5 image preview and that is free. How crazy is that? You can actually use an API for generating images completely for free. I've never seen that before. It's pretty amazing though. And then if you want to you can also go to Google AI Studio. You can have a look and see which apps are actually currently using it as well. So you can see here for example open router client root code etc. You can use these inside here. Really cool stuff. And then you could actually build your own image generation tool using the API. So for example, you could take the API requirements, go AI studio, code something out, and then build out your own SAS tool. For example, like a YouTube thumbnail generator for example. So, that's going to take a couple of minutes to generate the video. Here we go. We've got it ready to go. So, that is the image turned into a video. Let's play that again. Let's go full screen on this bad boy. Here we go. — There we go. That's actually got sound and commentary on the background, too. Plus, you can see the rain in it, too. Right. falling down from the sky in the background. The attention to detail is crazy. So, you can literally bring anything to life. AI code says, "When does Gemini 3. 0 come? " Who knows? But I would imagine that it's probably going to come. A lot of people were predicting August. So, it's pretty much the end of August. I'd expect it to come next month to be honest with GP5 coming out, Claude 4, etc. Gemini are due an upgrade when it comes to this as well. So pretty cool.

Now you can also get access to something called Whisk. And with Whisk, using the new Gemini 2. 5 Flash Banana update, you can now start using this as well. Right? So if you go to Whisk, Whis Labs, then you go to my library, you can create a new project here, right? So you can add an image, and then you got the subject, the scene, and the style, right? Which is really cool. And then Nyla says, "Any way to use it without so many what sort of policy restrictions are you facing? " Because I've not found any and people comment on the live stream there's policy restrictions and I'm like I haven't seen any problems with it. So I don't know like what people are creating with it that's so crazy that it refuses to generate. I've not seen any issues with the policy updates on Gemini at all so far. So let's have a look what we got here. So we can add the subject to scene and the style. All right. So if we go back here, what I can do, let's go back to the what we got here. I'm going to go back to Gemini. Let's say, okay, create a subject scene and style for my movie about a gangster cat. Something like that. Describe each visually plus separately in a short way. All right. And what we're going to do is just use this to generate out on Google Whisk. Now, if you haven't seen what you can do with Whisk, it's pretty cool. So, you add your subject, you add your subject again here, and then you add a scene, right? And so you can see, for example, they've taken the face from here, plugged it onto the character, the suit as well, and then the scene is the jellyfish ocean. So you can really like create like a full character scene using this process. It's absolutely crazy. So for example, here we've got the scene. So what I'm going to do is we're going to go with create images and I say create this image and that'll be the subject. Now, if we ask chat GPT to generate the same thing. Let's see which one is better. So, you can see how fast that was. That's looking pretty cool. The cool thing about this is it's similar to Photoshop, but it's AI style. So, you can and you can do this for free. So, for example, with this one, you can say add a baseball bat instead of the gun. And then, you know, you've got the image right here. You can see that there are some lack of details here. So, for example, it's not quite gripping the baseball bat. So, I'm going to say make sure the cat is gripping the baseball bat, plus make him look more like a gangster. So, it's quite easy to go back and forth in the chat. I haven't seen any limits on generations, but there must be a limit as well. So, we'll see what we get here. There we go. All right. So we're going to take that. Then we're going to go to whisk. We'll plug in the subject. Now we can generate the image directly here or we can drag and drop as well. So we can upload an image. Right? So if we upload an image like we can add that. If we want to add a scene, we can actually just take the text from the scene. So let's pull that over. Click generate. And then we've got the style, right? And so inside the style, I'm going to enter the text here. Hit generate. And there we go. So we've got all the styles and the scenes and everything else. So now what it should start doing is generating a video based on the subject, the scene, and the style. Now if you want to animate this, you can. So you can click on animate here. And then it's going to start generating this like so. Now if we want to compare the outputs of Gemini versus say chat GPT image generation. Let's take this one for example. We'll go straight onto the chat tools create images and side by side we'll compare chat GPT versus Gemini. So, we've got this one versus this one and we'll see which one generates the fastest. And you see how much quicker Gemini is, right? It's actually mind-blowing. So, if we compare these side by side, like GPT5 is still getting started. It's going to take a long time. It's actually it's not using GP5, I don't think, to generate the images. I think it'll be easy in there. The new API that just came out a few months ago from Chat GPT to generate that. But you can see how much quicker you get outputs. Right. So, now we've got the video with the sound. So if we just change the sound here. All right. And before we get into that, look at the difference in the quality of outputs. Right. This is the cat versus this one. And I would genuinely say this one is a lot nicer. Chat 55 seems to default to this kind of like almost drawing style image. which I don't think it's anywhere near as good as this one. And then if we go on to Whisk, Whisk has already generated the video based on our scene. So if we open this up, we've got the original image and then we animated that into a scene. And so if we go full screen on this, we actually have the sound in the background and it's generated in Veer 3 and the three different inputs we gave it. And there you go. Right. And then you can download that, use it how you want to, etc. And you can see the generations in the top that we have left. So it works really well with whisk. Honestly, was that like the most mind-blowing example I've ever seen? No. But it is really good. Juliano says, "Good morning, Julian, bro. Google throwing bombshells one after another. " Like it's just it's relentless. Constantly giving out new updates and new cool stuff is absolutely amazing to be fair. The fact that you can do all this for free as well is just mind-blowing, isn't it? Let's have a look at some of the other creations that people have made for inspiration on Twitter. I want to see what other people are making. So, you can like, for example, this is a pretty cool feature. You can take you can like edit someone's hair and just see, okay, what do they look like with different hair colors, if that makes sense. And then also you can see here for example you can edit a photo or a selfie and just add different background styles to it. Now on top of that you've got for example let's say you got a picture of your grandma and granddad like these guys and then you've got an alien and an ape at the end and then also you can turn a black and white image into something with color. So that's really cool what you can do here. You can see for example someone's changed the wedding photos to a potato because why wouldn't you? So this is basically it. Thiago says, "Why are you using whisk instead of flow? " Because flow whisk is just flow you can use as well, but I really like the idea of just having three different options here. And then I think that gives you the most out of the Gemini 2. 5 image generation. Whereas, for example, inside Flow, not that much has changed, right? It's pretty much the same, which is video generation. Ola says, "The real Julian, that's it. " And then Arian says, "Nano banana is really crazy. " I would agree with you. So let's go back here. Let's try and do something. So for example, if we take this image here and I say right now give the cat green hair and you can basically edit this like Photoshop except you're giving it textbased prompts, right? Instead of editing it manually. So, the need for designers is becoming less and less. You can see the hair right there. Pretty good. The only thing that I have found it struggles with, and I'll show you some examples in a minute. Some of the

stuff that I've seen it struggle with is actually editing thumbnails, YouTube thumbnails. So, if we go over here, like text replacement. And so, for example, here, this is something, you know, that is not good about it. So, I said, for example, change GP5 instead to Photoshop instead inside this image. and it just totally misspelled it as you can see right here. And you see how it's really sloppy with the way it's highlighted the background text. It should only be highlighting this word right here, but it's highlighted hop is and then as well doesn't make sense. And then I said Photoshop, not Photoshop. It redid it, but you see how sloppy that red background is. So with text editing, it doesn't seem to be great. And then I said change Photoshop is dead to nana banana. and it's got like a really big space inside here. So, even though it edited it how I wanted it, it didn't keep the design really nice. So, it's just something to bear in mind is like the text editing pretty trash.

Arian said, "Can you compare the output from AI Studio and Gemini of image generation? " That's a great one. So, let's do that. We're going to go into AI Studio here and we'll go into Gemini. And inside here, we're going to say, okay, generate an image of a woman marrying a potato. And then inside the settings, you just change this to featured. And then Gemini 2. 5 flash image preview. All right. So, click on that. And then we'll do the same inside Gemini over here. Create images and we'll see which one does the best. Now inside AI Studio, it does have a limited free quotation of the outputs. Okay, now this is interesting too. So it says model isn't available right now. Please wait a minute and try again. So, it's probably being flooded on the servers in terms of who's using it and that sort of thing. You can see you can actually use the old version here. Let's try it again. I'm going to give it one more world and then if it doesn't work, we may give up or we may try inside open router. There we go. So, it worked here. One thing I noticed inside Gemini 2. 5 Flash inside Gemini directly is that the images seem to be really cartoonike pretty much every time. Whereas for example, if we have a look here, this is more slightly more photorealistic. Which one do you prefer, by the way? Let me know in the comments. Anyone who's watching, do you prefer the AI Studio output or output from Gemini? I think if I had to choose one, I'm probably going to go with that one just because it's less cartoony and it's more photorealistic, which is what I think we all crave from photos really. But either way, pretty impressive right there. So, you can see an example right here from Ting. Ting is tweeted and said, "Give me red hair. " And basically, they got red hair right there. Let's try that out now. All right. So, let's take this image here. There we go. And then we're going to say, okay, make my hair long, crazy, and bright red. We'll see what that does. Arian says, AI Studio output similar to what we have seen in LM Arena. And there we go. Look at this. Cool. I might get my hair like that. Why not? Let's say make it bright green dreadlocks. It is crazy easy to just edit the photo, right? Yeah, that's beautiful. That's what I'm talking about. Look at that hair. You may see me on the next live stream with this hair. I can and cannot confirm. So, that was pretty easy to do. The other thing that I heard is that it's actually coming out to notebook LM soon too, which would be pretty cool. Let's have a look through the recommendations as well for this. It says you want the subject, the composition, so how it's framed, action, location, style, and editing instructions. All right. So, let's take the prompting guidelines from Google here. We'll plug that in here. Then, we're going to say

okay, create a funny image prompt for a cat eating bananas. here. Whatever you do, make sure it's not cartoon style. So, we'll take that and I'll say this is based on the guidelines I've given you. There we go. All right. So, we got the subject, composition, action, location, etc. Let's go back to Gemini now. And then whilst we're waiting for that to generate, let's go back here. So, subject, who or what is in the image, composition, how is it framed? So like closeup or wide, what's the action that's happening? Location, style, and editing instructions. And then you can preserve character appearances. So for example, it can maintain the likeness. You can see side by side how they look. And now we've got an image. And yeah, way better, isn't it? way more photorealistic, much more interesting. That's what we want. All right. Now, I'm going to go back to prompting wizard here, and I'm going to say, okay, actually, yeah, how can we make this funnier? And then it asks, do you want to upgrade the prompt? There we go. All right. Boom. Now, this is 10 times better for getting better outputs because it's using Google's guidelines. So, let's take that. We'll plug that in. In the meantime, what I'm going to do is I'm going to create a custom GPT for generating outputs with this customuilt prompt trainer for Gemini. So, I'm going to say, okay, give me some instructions for a custom GBT to generate high quality images every time. the Gemini 2. 5 flash based on the custom prompting guidelines from Google. All right. And then if you want to create something like this, what I'm going to do is I'll go to GPTs create image prompter for the banana. And then if we go to here, we've got the custom instructions. There we go. Now, we're going to create this custom GPT. And if you want access to that, I'll put it inside the AR profit boardroom. just go to the SPS and then you'll see the custom GPT for generating better images here and that's just custom built so that you get a better tool for generating the outputs because what I found is if you go directly to it it's not particularly good as you've seen today but if you use the custom GPT way better right so you can get that inside the AR profit boardroom link in the comments description if you want it so there we Go. Now we're talking way better. Banana eating competition. We got a monkey looking angry. We got a cat with a bib and a bandana. What we can do from that is now we can say, "Okay, turn this into a video. " So if I download that and you see it's got little banana reward there, too. So we've downloaded that. Then we're going to go to new chat here. Take that image and turn it into a video. And that's one of the good things about Veo 3 recently is it's got this really cool image uploader. So you can have like character consistency when you're generating those videos. So if I say create a video from this cat banana eating contest, then it should turn that into a video. Let's see what else we got here. Make targeted transformations with precision. So high quality. Three different images here, right? Let's have a look at this. So, high quality photo of a modern minimalist. Living room with a gray sofa. Editing. Change the sofa color to a deep navy blue. So, you can see that's done there. And then prompt three. Now, add a stack of three books to the coffee table. And you can see them there. Right. So, tiny little edits like this are really cool. So, you can see here we've got a gray sofa. Then, we've got exactly the same image, but a blue sofa. And then three books on the table. Boom. Let's generate the video. Let's come back over here. So you can fuse two or my ideas into a single striking image. Interesting. So you can generate a photorealistic picture of an astronaut, a picture of an overgrown basketball court, and then upload both and combine. Show the astronaut dunking a basketball in the court. So this image plus image, this image equals this image right here. So you can I think it seems like you can add multiple images if that makes sense. By the way, we've got the video back now. Let's have a look what we got. Should have sound on there, too. Rafie says, "I like banana. " Good for you. So, that's winning the banana race there. pretty simple and easy to do. Yes. So, you can create images like so easily using this process. Let's see if we add like the last two images here. We'll see if we can combine them. So, we'll create a jungle. I'm actually I'm going to use the custom GPT to generate something better. So, we go back to this one. Click on that. So, it gives us the prompt, which is beautiful. Let's take that to go into Gemini. That was so easy to generate the prompt, right? And that should give us much better outputs. Loving this custom GPT. So now we've got the image of a jungle. I'm going to download that. Right now I'm going to say, okay, funny basketball playing Bigfoot, something like that. So now we got this prompt as well. So

we're going to generate two separate images. One of a jungle and one of this one. So there we go. So now we've got both images here. We should be able to combine them. So we've combined both these images, the jungle and the basketball of Bigfoot. And then boom shakalaka. There he is. That's got the background of the jungle. Some parrots on the basketball hoop and Squatch on the basketball kit. Pretty cool. And then you can also change like the style of the image to be like drawing mode or what would happen if something happened, right? So for example, okay, if this woman dropped the cake, what would it look like? That's an example. And you have to use the same session if you want the context. So, if we go on the same session here, we say generate an image showing what would happen if Bigfoot did a slam dunk. And there we go. Boom. Big foot in the jungle doing the slam dunk. So that's basically it. You know how to edit

images, change them, create outputs, create videos from them, use Google Whisk, do all this for free, use the API inside AI Studio, we have covered a lot there, plus how to custom prompt Gemini to get better outputs, too. If you want everything that I've talked about today, including the custom GBT and the free tool that I built for you for creating better images, get that inside the AI profit boardroom link in the comments description. You can see we update this group daily with new updates, tests, tools, etc. Also, inside here, you can post any questions you have. It's a really active community of people serious about AI and growing their business and scaling and automating with AI. On top of that, we have loads of different automations inside the classroom, including tons of different NA10 automations. As you can see, we also have for example a six week master class on AI crash course and all the automations I use in my business Q& A call recordings plus inside the calendar here you can see that we actually do five live coaching calls uh a week right basically what that's designed for is number one to create a community feel where everyone can just share and learn together but number two we have experts running those sessions so you can ask like a PhD scientist any questions you have about AI and the goal is just to make you save time, help you grow your business, and automate whatever you want with AI. So, feel free to get that link in the comments description. Appreciate you watching. Um, if you want us to automate everything for you, feel free to book in a free AI automation session on that call. Basically, what we're going to do is we're going to look at where you spend your time. Then, we're going to figure out what automations you need in your business. And from there, if you want to become a client, then we'll just go off and set it up for you. So, thanks very much for watching. I appreciate it. and I'll see you on the next one. Cheers. Bye-bye.