Nano Banana2 incoming! The dust seems to have settled, and the first big things are on the horizon. Following the huge success of Nano banana, the second iteration is about to be released.
TestingCatalog News 🗞
 @testingcatalog
Nov 5
BREAKING 🚨: Google is preparing Nano Banana 2, "GEMPIX2," for an upcoming release! A new announcement has been recently added to the Gemini website, which means that we should expect a release within the following weeks.

Nov 5, 2025 · 1:25 AM UTC

27
48
7
974
Replying to @kimmonismus
@tchinnom @newdtfullowning @oliver__99_ Nano Banana 2세대가 몇 주 안에 공개된다고 하네요 👀❗️
2
1
6
Replying to @kimmonismus
Looks like we're in for another wild ride. Hope the banana peels are less slippery this time! 🍌
Replying to @kimmonismus
"Edit this image" "Okay, here you go" "That is the exact same image"
35
Replying to @kimmonismus
Damn. hard to see how you make a version 2 of that 🤣 It's so good for editing. I guess we'll see!
Replying to @kimmonismus
Actually is Nano Bannana 3
Replying to @kimmonismus
Nano banana 2 before gta 6
Replying to @kimmonismus
super excite!
Replying to @kimmonismus
The second will be amazing, the first one is already excellent
2
Replying to @kimmonismus
Clearly, Google’s about to peel back another layer of AI surprises.
Replying to @kimmonismus
Excellent. I hope they will continue reducing censorship.
1
Replying to @kimmonismus
Woah!
1
Replying to @kimmonismus
That's another number one.
Replying to @kimmonismus
HUGE for my apps lol
Replying to @kimmonismus
Nano Banana2 might just take the momentum even further — if it builds on the first version’s hype and innovation, we’re in for something big.
Replying to @kimmonismus
SRKDAN
 @SRKDAN
Nov 5
🚨 NanoBanana 2 is coming… and OpenAI’s still quiet. Google’s new image model is reportedly just around the corner and it’s looking fast. Like 10-seconds-per-image fast. Meanwhile, OpenAI’s last major image model drop was almost half a year ago, and that one still takes roughly 15 seconds per render. Sure, OpenAI’s output feels more “modern” or “stylized,” but NanoBanana’s edge? Accuracy. The results are closer to what you actually prompt for. If Google really nails this next iteration, it could set a new benchmark and honestly, I think it’s time for OpenAI to respond. The creative AI race is heating up again, and the gap’s starting to show.
Replying to @kimmonismus
Yessss! We need a updated Nano Banana, even though it’s a great model I would love something even stronger 🙏
Replying to @kimmonismus
Wow!!
Replying to @kimmonismus
Will it understand when I say “I need in landscape NOT A SQUARE” or will that be still too complex?
1
Replying to @kimmonismus
Exciting times ahead with the next chapter of innovation.
Replying to @kimmonismus
I hope they fix the pixel problems when there images hits a real bitmap! I have to upscale and use camera raw to save its pixels from breaking!
Replying to @kimmonismus
I just hope they fix prompt adherence, as nano banana is really good when follows your prompts, but like 20 percent of the time, the prompt is not followed.
1
Replying to @kimmonismus
Distraction from G3
1
Replying to @kimmonismus
Micro Banana? Medium Banana? GIGA BANANA?
1
Replying to @kimmonismus
If they learned from the first version this could actually be a game changer for edge computing on mobile devices
3
This tweet is unavailable
Replying to @kimmonismus
Interesting to see a second iteration already.
Replying to @kimmonismus
Woahh!!
Load more