Google Launches Nano Banana Pro AI Image Generator With Improved Text Capabilities

[MacRumors]

Google today announced the launch of Nano Banana Pro, an image generation model that is a followup to the Nano Banana model that came out earlier this year. Nano Banana Pro is built on Gemini 3 Pro, Google's latest and most advanced AI model.

Google says that Nano Banana Pro is the best model for generating images with correctly rendered and legible text, from a short tagline to a long paragraph. Nano Banana Pro's text expertise allows for the creation of mockups and posters with different fonts, textures, and word style, plus it works in multiple languages.

Content is more helpful than before, with Nano Banana Pro able to create text-heavy infographics and diagrams for learning purposes. It can connect to Google Search to gather information for a graphic, and blend more elements than before. It supports uploading up to 14 images and can maintain the consistency and resemblance of up to five people.

There are image editing tools in Nano Banana Pro that can be used to select, refine, and transform parts of images with localized editing. Camera angles, focus, color, and lighting can be adjusted, and there are 2K and 4K resolution output options.

With the launch of Nano Banana Pro, Google is giving Gemini the ability to tell if an image was generated with Google AI. AI-created images feature SynthIDs and a visible watermark on images generated by free and AI Pro tier users.

Nano Banana Pro is available in the Gemini app when using the Create images feature with the Thinking model. Free users will receive limited quotas, and then will revert to the original Nano Banana model. Google AI Plus, Pro, and Ultra subscribers can generate more images before hitting the limit. Nano Banana Pro is also available for Google AI Pro and Ultra subscribers in AI mode in Search.

Google is making Nano Banana Pro available for companies that use Google Ads, and developers can use it through the Gemini API and Google AI Studio.

Article Link: Google Launches Nano Banana Pro AI Image Generator With Improved Text Capabilities

 

[orbital~debris]

How do companies… Google… talk about products they name in such a fashion with a straight face?

 

[CharlesShaw]

The 2019 song Nana Banana is about to get some more listeners.

 

[neuropsychguy]

I just tried it on my default test -- trying to have the generator accurately render something medical-related. It did an impressive job on the first try. It wasn't perfect but was the best result I've seen yet with any tool. The second try was nearly perfect.

Edit: With my particular test Grok was pretty terrible. ChatGPT was better than Grok but not as good as Gemini (and I needed to modify my prompt somewhat to get it closer to what Gemini produced). Google went from behind to being very comparable to or ahead of the others. I'm rethinking my ChatGPT subscription at this point and switching mostly to Gemini.

 

[edvj]

How do companies… Google… talk about products they name in such a fashion with a straight face?

The hilarious, 2:30 AM origin story of how ‘Nano Banana’ got its name

 

[tliede20]

I've been trying all morning to get it to render stuff, but there must be a lot of congestion because I continue to get error 13.

 

[kiensoy]

No need for this. I have Image Playground.

 

[FangRock]

Just tried it. It’s incredible. Shows how far behind Tim’s Apples are.

 

[Porco]

Presumably Micro Kumquat, Mini Melon and Tiny Satsuma are all now playing catch-up, and don’t know whether to go Max, Plus or Ultra.

 

[CATiNTHESTORE]

Someone in Google plays to much Overwatch 2 )))

 

[sniffies]

How does it compare to Midjourney and other early AI models?

 

[svish]

Will try this out. Hoping to see some of these capabilities integrated into Apple software in the future.

 

[delsoul]

Honestly, I’m AI’ed out. Really just tired of it. Everyday is some “new and shiny” release, really lost its luster and appeal.

 

[Jumpthesnark]

Thanks Google, for stealing from copyrighted material found on the internet created by photographers, to make it easier to put photographers out of business.

Eventually, the only images left for the AI to comb through will be the AI slop itself.

 

[bluecoast]

Next week Apple will respond with Genmoji+ - a revolutionary way to create even more compelling emojis - at only $4.99 per month!

 

[MRrainer]

Who needs this? What problem does this solve?

All the GPU cycles wasted for...what?

 

[Diopter]

Google engineers be like: "mmmm, industrial-scale copyright infringement for profit 🤤"

 

[amartinez1660]

Honestly, I’m AI’ed out. Really just tired of it. Everyday is some “new and shiny” release, really lost its luster and appeal.

A good trick to this is to just let the articles and news pass by. I’m AI’ed too, but I looked for this article, clicked on it, started reading the comments.

Can just apply the same to when someone is Pizza’ed out or French Fried out: don’t order pizza or fries, don’t even browse for pizza or fries, heck, not even the “50% off today on Domino’s!” ads are registered by the brain.

 

[ginkobiloba]

Fantastic. I can’t wait to see the new level of totally undetectable high-quality fake news, disinformation and scamming this will unleash.

Very useful progress !

( “Apple is so behind the times” posts incoming in 3..2..1..)

 

[iMac The Knife]

But, does this come with the Fee fi fo-nana add on? That's the real question.

 

[musteky]

Meanwhile Apple is going completely banana with its own AI

 

[The Game 161]

Just tried it. It’s incredible. Shows how far behind Tim’s Apples are.

Easy to see why Apple are adding Gemini model now

 

[richinaus]

Thanks Google, for stealing from copyrighted material found on the internet created by photographers, to make it easier to put photographers out of business.

Eventually, the only images left for the AI to comb through will be the AI slop itself.

maybe rephrase this to 'average' photographers [and every other average 'creative'].
I dont see in the near future AI replacing the actually talented people who have a style, taste and a unique eye. This applies to music, art, photography, architecture and every other creative industry.

Unfortunately the average vastly outweigh the talented in terms of numbers of people.

 

[Jumpthesnark]

maybe rephrase this to 'average' photographers [and every other average 'creative'].
I dont see in the near future AI replacing the actually talented people who have a style, taste and a unique eye. This applies to music, art, photography, architecture and every other creative industry.

Unfortunately the average vastly outweigh the talented in terms of numbers of people.

Not going to rephrase it at all. The people you call "average" deserve work too, and AI tools like Nano Banana Pro are being marketed (just watch the ad) to the would-be clients of all kinds of photographers and other creators.

Not every client has the resources to hire top-tier photographers, musicians, etc., and there are plenty of creators who don't command high day rates. They still deserve to make a living without AI taking it away from them.

 

[JMalone]

Not going to rephrase it at all. The people you call "average" deserve work too, and AI tools like Nano Banana Pro are being marketed (just watch the ad) to the would-be clients of all kinds of photographers and other creators.

Not every client has the resources to hire top-tier photographers, musicians, etc., and there are plenty of creators who don't command high day rates. They still deserve to make a living without AI taking it away from them.

We need to point out the obvious to those who hate working people.

1. 99% of photography jobs are low paid overworked junior photographers not some rich model humper in Epstein’s black book.

2. Nearly all photography jobs involve taking photos of real product or real events. Generative AI models have no idea what those products look like until you train the models on the photos, which means the photos have to be taken in the first place. No photos - no images for AI to feed on.

3. No, you can’t train the models on CAD files. CAD files and text files cannot accurately describe every physical property of a product. Even product designers don’t know how a product will look like from designs alone and have to build iterations of product until they are happy.

4. Tools like Nano Banano are for those mostly unknown companies shilling fake products and buying scam ads. Google is selling them to the tool to generate the ads and then selling them the ad space. Platforms like Meta/Facebook also profit from this scam economy and last week they admitted they earn several billion for scams.

Meta is earning a fortune on a deluge of fraudulent ads, documents show

Meta projected 10% of its 2024 revenue would come from ads for scams and banned goods, and it internally estimates that its platforms show users 15 billion scam ads a day, company documents show.

www.reuters.com