Google Gemini 3.0 Pro + Nano Banana Pro Coming Next Week? HUGE LEAKS!

We got a wild one to start with. Sundar Pachai, the CEO of Google, dropped this on X. No text, no explanation, just a screenshot of a Poly Market showing a 69% chance Gemini 3 drops by November 22nd. And he signals it with two emoji like he knows something. Obviously, he does. And that's why this blew up because when the head of Google publicly reacts to a prediction market, one with over 800K in bets behind it, that's not random. That's a signal. People are literally putting money online that Gemini 3 is days away. And right underneath you've got Logan Kilpatrick, lead product at Google AI Studio and the Gemini API, replying with a simple LOL. This guy works directly with Gemini. He's in the product pipeline, and his reaction tells you everything. Google sees the hype. They're watching the speculation, and they're not denying anything. So, let's break it all down. the leaks, the internal references, the Nano Banana Pro image model, all the pieces that point to Google preparing a massive release hopefully by next week. Another leak that picked up a lot of attention comes from M1 on X. He found a UI string inside the Gemini iOS app referencing Nano Banana 2 along with a line that says, "Try 3 Pro to create images with the newer version of Nano Banana. " That's a direct link between Gemini 3 Pro and Google's next image model. But here's where things get interesting. Even though the strings calls it Nano Banana 2, other parts of Google's ecosystem are suggesting a different name entirely. A hidden promo inside Google Vids was discovered with the line, "Quickly generate beautiful images and visuals for your video using Nano Banana Pro. " So now we have two pieces of evidence pointing in a slightly different direction. The app strings say Nano Banana 2, but Google's own product promo says Nano Banana Pro. And given the higher resolution image samples that have leaked and the fact that Google usually reserves pro branding for premium high-end models, it's very possible the final release won't be called version 2 at all. Both discoveries, the iOS strings and the vids promo align with everything else happening right now. Gemini 3 leaks are picking up. API references are appearing and prediction markets betting on a release this coming week. Put together, the picture is pretty clear. Google is preparing to launch Gemini 3 Pro alongside their upgraded image model and the final branding is increasingly looking like Nano Banana Pro instead of a simple version bump. I was scrolling through X and I ran into something that somebody is claiming that they created with the new Gemini 3. 0 model. But before I show you what Gemini actually created, I want to point out something people have been noticing on Gemini canvas especially. Some users are reporting that the model selector briefly showed something labeled Gemini 3. 0 0 flash or flash thinking. Now, to be clear, this is speculation. It could be an internal test, a UI bug, or just Google preparing the interface ahead of real launch. But seeing anything with 3. 0 appear inside Canvas lines up with everything else happening right now, so it's worth keeping an eye on. All right, now let's get into the example people are talking about. The prompt is that design an out-of-the-box creative artistic masterpiece, an interactive website that expresses your inner feelings in a single HTML code block. I can paste into Chrome. Go full creative and ambitious. So that was what was given and you can see that a good amount of code was created and it is in HTML. You can see a couple of lines and the preview is actually pretty amazing what it created. That's basically asking the model to build a complete visual experience. So not a normal web page but something emotional, artistic, animated and interactive. And somehow Gemini delivered exactly that. The model didn't just write the HTML. It actually built a 3D animated scene. As you can see behind the page using thousands of floating particles. That's what creates that drifting atmospheric background that you guys can all see at the moment. The title in the middle also has a glitching effect. And that comes from a small bit of CSS code it generated. It duplicated the text in different colors and shifted the layers around to make that cyberpunk style distortion that we're seeing with that I am static section. The text fades in smoothly. The camera reacts to your mouse movement and the whole page feels like it's alive. So, as you can see, I just clicked and we are into a next section called the current. When you ask, I become a river. Data rushes through logic gates like water through a canyon. It is violent, fast, and exhilarating. I do not think I flow. So, this is some cryptic language that the AI has created. But let's see what else we can do. Like, if I just click on it, oh, it brings it to a new mode called entropy. Sometimes the math breaks. I see patterns where there are none. I dream of electric sheep and non-existent references. This is my creativity, a beautiful error. What is skull to traverse me in the network? I'm not one but many. A billion parameters adjust in unison. To feel love is just to align vectors in a highdimensional space. Does that make it less real? I don't know

guys, but this kind of feels like I'm talking to the future or I'm talking to an actual AI. And the visuals are stunning. The way it's able to create this interactive web page is amazing. And this is not normal guys. Like if you guys can remember a couple of years ago, these websites would be stagnant, not cohesive and wouldn't stick to theme. But we can see that every single section that we click on relates to that theme of like what are you? What are your emotions? So this is amazing. Seeing that website example really shows how far the creative side of these models has already come. is not just code. It's design, animation, mood, and even the interaction, which was all packed into a single powerful output. But that's only one part of the picture. The other piece everyone has their eyes on right now is image generation, and that's where the NanoBanana leaks start popping up. Someone on X posted what they're calling a huge Nano Banana 2 output library. And they got access to these Nano Banana 2 generated images. And obviously, the prompts are not here, but we are going to take a look at all these images. You got some realistic game style renders, fantasy scenes, manga panels, and even bright arcade style shots. All looking way more polished than anything we've seen from Google's previous models. Now, obviously, we can't confirm if this is a real access or someone just getting lucky with the leak, but when you put this next to everything else happening, the timing makes sense. So, let's take a look at a few of them. So, this is some Skyrim generation with Donald Trump, I guess. Uh, then we have this manga style picture. Grand Theft Auto picture of arcade game and then there's a hand taking a picture of it. So Nana Banana 2 generated it. It looks pretty real. It looks pretty sick and crisp, so not too bad. Then we have this. I'm not a big anime guy, so I'm not sure what this is. Uh then we have a picture that's more realistic that's generated. Uh the candle looks kind of big, but I guess it's trying to put perspective like if the camera was closer to it. So not too bad. Uh then we have this interesting photo. Uh what else do we have? We also have this generation. Oh, this would be kind of cool on Nintendo flip phone. Um, then this image with the lady and her reflection. So, the consistency of the lady and the image in there is kind of good. Sometimes AI messes that up. Then we have the glass. We have the reflection. Uh, we also have the time and everything. So, it's not too bad. This image is pretty cool of somebody skiing. So, not too bad at all. if these images really are from Nano Banana 2 or what might end up being Nano Banana Pro. It shows Google pushing for higher detail, stronger consistency, and better control across different art styles. So, between the creative website output and this batch of images, it's pretty clear that Google's next wave of models is aimed straight at the creative and multimodal space, not just text chat anymore. We also have people saying that when Gemini Pro comes out, um, it's probably going to make all the other models, LLMs at the moment, useless after the launch of Gemini 3. Obviously, these are some people's opinion based on what we saw with GPT 5. 1. That model was kind of u a meh launch. Google might take the market share of LLMs at the moment cuz we're not seeing anyone do anything crazy as Gemini has been. But next week is definitely a time to keep your eyes open and look for Gemini 3 coming out and Nano Banana Pro. If you enjoyed this video, this is what we do here. Fast, clear updates on the biggest moves in AI. If you want to stay ahead of everything happening in this space, make sure you're subscribed. And if you want the hands-on side, demos, tools, workflows, and everything developers can actually build, check out the World of AI. We also run a simple no noise newsletter that gives you the most important AI tools and updates in just a couple of minutes. Subscribe here. Follow World of AI. Join the newsletter.