Google launches Nano Banana 2
Google launches the next version of their image generation AI model Nano Banana 2, and it's already live in the Gemini app. The model is basically the Gemini 3.1 Flash Image model which has improved image generation capabilities along with improved speed. This launch post from Sundar Pichai has some additional details with a demo video as well.
I can access it via the Gemini app, but it asks me to add a paid API key to access the new model inside Google AI Studio, as you see above.
For the new model, they have introduced multiple new aspect ratio (4:1, 1:4, 8:1, and 1:8), and thinking level (minimal and high) as well. And from a few image generations I did:
- the model is faster
- image quality looks better
- does fewer errors, and
- does texts really well
I am still using it and will keep updating this post with new learnings.
Update:
I just came across these Nano Banana 2 image generations, and they are amazing.
Webmentions