Nano Banana 2.1 is now on Leonardo.Ai
Releases | Published on | Last updated on
2 min
Nano Banana 2.1 is Google's image model for consistent characters and products, and it's now in the Leonardo.Ai image generator. Add your people and your product as image references, up to 14 in all, plus one video. Then put them in new scenes where they still look like themselves.
Keep your cast consistent
Add a portrait of each person as an image reference and say where they go next. For the two scenes below, five portraits went in. All five people turn up in both, packed into a lift and then lined up at a bakery window, and everyone keeps their own face, hair and clothes.
Five reference portraits in, the same five people in a lift and at a bakery window.
Put your product in every shot
Add a photo of your product as an image reference, then place it anywhere you can picture: on a fishmonger's ice, on a street poster or in a scene with your cast. Your product still looks like itself in every shot.
One product photo in, the same bottle on the ice and on a row of posters.
Combine up to 14 references in one image
Use up to 14 image references in one generation. This one uses all 14: a cinder block, a teapot, a bowling ball, an egg and ten more everyday things. Say how they fit together, and you get one image that uses every piece.
Fourteen references in, one tower out.
Make a still from your footage
Add a video as a reference and make a still that matches it: a thumbnail, a poster or key art for a campaign. You can add one video to each generation.
A frame from the reference video, and the gig poster made from it.
What you get
- Up to 14 image references and one video reference in each generation
- Up to 8 images from one prompt
- Ten aspect ratios, from 9:16 to 21:9
- 1K, 2K or 4K output, with the token cost shown before you generate
On the free plan, you make one image at a time with one image reference, and your images are public.
Start your first scene
Say who's in the shot, where they are and what each reference brings to the scene. Prompt Enhance turns off when you add an image or video reference, so the scene comes from the prompt you write. The Nano Banana prompt guide has examples to borrow.
See more of what you can make with Nano Banana 2.1, and how it compares with Nano Banana 2 and Nano Banana Pro.
RelatedArticles
Leonardo’s Pro Upscaler: Built for AI-Generated Images
Veo 3 vs Sora 2: When to Use Which Model
Google’s Veo 3 and OpenAI’s Sora 2 have become so powerful they’ve shifted the debate for creators. The question is no longer “Which model is better?” but “Which model is better for this specific task?” In this guide, we find the answer. We will run both models through the same prompts to directly compare their […]
Nano Banana – aka Google’s Gemini 2.5 Flash Image – Arrives On Leonardo.Ai
Human creativity is equal parts craft and imagination – and the tools we use should feel just as dynamic. Enter Nano Banana, now live on Leonardo. Built for creators who demand precision without sacrificing artistry, this model puts an entirely new level of control at your fingertips. Whether you’re sketching out a concept from scratch […]
Lucid Origin – Leonardo.AI’s latest Gen AI model.
Setting a new bar for visual generation Generative AI is evolving fast – and Lucid Origin from Leonardo.AI is here to raise the standard. With great diversity, more dynamic range, and trained for Full HD renders, Lucid Origin sets a new benchmark for aesthetic fidelity. Built to impress, it breaks through many common limitations of […]