We made Nano Banana 2 (Gemini 3.1 Flash Image) the default model in Creativly Flow shortly after Google shipped it. We ran it through production prompts, and it held up.
What Changed in Real Workflows
The biggest change was iteration speed. We got to a usable first draft faster, and prompt adherence felt more reliable across both AI image generation and image editing.
Google positions Nano Banana 2 as Pro-level capability at Flash speed. Our own tests matched that direction. You prompt, you see. No more waiting around between ideas.
"This is the first image model where waiting stopped being part of the workflow."
Text Rendering and Translation
Google highlights precision text rendering, translation, and stronger instruction following in Nano Banana 2. In practice, this matters most for social graphics, mock ads, greeting cards, and signage concepts where readable text is part of the deliverable.
Google also calls out output up to 4K and aspect-ratio controls. On Creativly, those options depend on the provider route selected for the run.
Editing That Preserves Intent
We saw the biggest lift in reference-driven edits where composition had to stay intact. According to Google, Nano Banana 2 supports subject consistency for up to five characters and fidelity for up to 14 objects, and that matches the direction we observed in edit-heavy workflows.
ORIGINAL
EDITEDGrounding and World Knowledge
Google describes stronger grounding via Gemini's world knowledge and real-time web information. In our tests, prompts for specific landmarks were more reliable. Ask for the Eiffel Tower at dusk, and you're more likely to get the actual landmark details instead of a generic tower.
Available Now on Creativly
Nano Banana 2 is live now in Creativly Flow and the Focused Editor. New image nodes in Flow default to it, and you can still switch to other models at any time from the picker.
It supports both generation and editing workflows and works with platform credits or BYOK.