Nano Banana goes Pro
• public
•
2 min read
Google just released the next version of it's much praised image model Nano Banana. Here's the quick run-down:
Key new capabilities
- Higher‑quality output, with support for 2K and 4K images instead of the original model’s 1024×1024 limit.
- Much better control over camera angle, lighting, depth of field, focus, and color grading for more “studio‑grade” results.
- Stronger text rendering so it can produce clean, accurate multilingual text in posters, menus, infographics, and slide designs.
Composition and consistency
- Can blend or combine up to around 14 images/objects into one composition while preserving detail.
- Maintains consistent appearance for up to about five people/characters across images, useful for branding, storyboards, and ads.
- Handles things like product mockups, style‑consistent variants, and turning rough sketches or doodles into polished visuals.
Time to revisit my old weavy.ai boards and plugin the new model to see the differences. Most of my applications are centered around product placement with as much detail and preservation of the original brand.
For most of those shots Nano Banana Pro did a better job than the first one, but in one instance (the bottle) the outcome was actually quite disappointing. But the older one is still around and that's the beauty of a tool like weavy.ai (now aquired by Figma), you can just iterate until you find the best model for your application.
Check out the weavy board if you want to look at the results in more detail.
