According to Ars Technica, Google has launched Nano Banana Pro, an upgraded version of its meme-friendly image generation model that’s part of the new Gemini 3 Pro platform. The model can now generate entire usable infographics with legible text in a single shot, blend up to 14 images while maintaining the appearance of five people consistently, and render creations at up to 4K resolution. It’s available to everyone in the Gemini app immediately, but free users will quickly hit usage limits before being downgraded to the non-pro version. Google is also adding AI image detection to the Gemini app, letting users upload images and ask “Is this AI?” to check if they’re Google-generated. However, the company removed visible watermarks for AI Ultra subscribers while keeping invisible SynthID detection, making it harder for humans to spot AI images.
Google actually wants you to use this
Here’s the thing about Nano Banana Pro – it’s Google finally admitting that AI image generation needs to be actually useful, not just meme-worthy. The original Nano Banana felt like a playful experiment, but this upgrade screams “professional tool.” Being able to create entire infographics without weird text artifacts? That’s huge for content creators and marketers who’ve been struggling with AI’s text generation limitations. And maintaining consistency across multiple people in images? That’s been a major pain point for anyone trying to create cohesive visual content.
This is about locking in the pro crowd
Look at the tier structure they’ve set up. Free users get a taste but hit limits quickly. Pro users get more access. AI Ultra subscribers get the highest limits and no visible watermarks. This is classic freemium strategy, but applied to AI image generation. Google knows that once professionals integrate a tool into their workflow, they’ll pay to keep using it. And removing watermarks for paying customers? That’s directly targeting the “I need this for client work” crowd who can’t have obvious AI markers on their deliverables. Smart move, honestly.
The AI detection game is getting weird
So Google is simultaneously making it easier and harder to detect AI images. The Gemini app can now identify Google-generated content through SynthID, but they’re removing visible watermarks for paying customers. Basically, they’re saying “trust us, we’ll help you detect AI – unless you’re paying us not to.” It’s a strange balancing act between transparency and commercial interests. And let’s be real – how many people are actually going to upload every suspicious image to check if it’s AI? The detection tools need to be built into platforms, not separate apps.
Where this fits in the AI arms race
Nano Banana Pro feels like Google’s answer to Midjourney and DALL-E 3, but with that distinctive Google twist of being both playful and enterprise-focused. The ability to make precise edits to existing images – changing camera angles, lighting, color grading without messing up everything else – that’s where the real value lies. It’s not just about generating from scratch anymore. It’s about iterative refinement, which is how most creative work actually happens. The question is whether Google can convince people to use Gemini for serious work when they’ve built their workflows around other tools. The improved resolution options and professional features suggest they’re serious about competing beyond just the hobbyist market.
