Readme
Nano Banana 2 is Google’s latest image generation model, built on Gemini 3.1 Flash Image. It’s the high-efficiency counterpart to Nano Banana Pro — optimized for speed and high-volume use cases while still producing high-fidelity images.
Google announced Nano Banana 2 on February 26, 2026, describing it as their “best image generation and editing model” — combining Pro-level visual quality with Flash-level speed and pricing.
What it can do
Generate images from text. Describe what you want and the model creates it — photorealistic scenes, illustrations, product mockups, whatever you need.
Render text accurately. One of the standout improvements over previous Flash image models — text in images is crisp and readable, with support for multiple languages.
Edit existing images. Pass in one or more images along with a text prompt to transform them — change backgrounds, swap colors, adjust styles, or combine multiple images into one scene.
Use up to 14 reference images. Feed in multiple images for style transfer, image combination, or complex editing tasks that draw on several visual references at once.
Key improvements over the original Nano Banana
- Higher fidelity output with richer textures and sharper details
- Much better text rendering and multilingual support
- Stronger instruction following for complex prompts
- New aspect ratios: 1:4, 4:1, 1:8, and 8:1 (in addition to the standard set)
- Multiple output resolutions: 512px, 1K, 2K, and 4K
Aspect ratios
Supports 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, and 21:9. Set match_input_image to automatically match the aspect ratio of your input image.
Output format
Choose between jpg (default) and png.
Links
You can try Nano Banana 2 on the Replicate playground.