"Use the product in the reference images. Place it [scene], [lighting]. Keep its shape and label exactly."
Up to 14 references in one request: name what must stay the same, and it holds it across the set.
Nano Banana — Google Gemini 2.5 Flash Image, the first of the Nano Banana family. Fast generation and editing that takes up to 14 reference images in one request, for keeping a product, a face or a style across a set.
Our two test pictures at 1K (the first two below) came back in 27 s and 29 s. The label has the right words, MORNING ROAST · 250 g, but wrapped onto two lines. It added an air valve and loose coffee beans we did not ask for. The portrait has rain drops all over the yellow hood and coat; the harbour is only blurred shapes. Two prompts are a small test.










Example prompts stay in English, which is what the models handle best.
"Use the product in the reference images. Place it [scene], [lighting]. Keep its shape and label exactly."
Up to 14 references in one request: name what must stay the same, and it holds it across the set.
"Change only [element] to [new value]. Keep everything else the same."
Edit-in-place works best when the change is named narrowly.
It is the first model of the family, Gemini 2.5 Flash Image. Nano Banana 2 and 2.1 are newer generations. It takes up to 14 reference images in one request.
It is not the model to choose for that. For posters, packaging or anything where the words must be right, use GPT Image 2.5, the default image model.
When your request plays to Nano Banana's strengths. Name it in your message or pick it in the composer; FlowPix never switches models on its own, and uses its default image model when you choose none.
Yes. The model picker in any chat lets you pin per-turn or per-session. Set as default in /settings.