Nano Banana 2
Gemini 3.1 Flash Image fast high quality AI image generation and editing
Nano Banana 2 (officially known as Gemini 3.1 Flash Image) is Google’s upgraded AI image generation and editing model that brings advanced visual creation capabilities to a broad audience. It generates detailed, expressive images from text and image prompts with sharp details, richer lighting, and improved adherence to complex instructions. Nano Banana 2 also supports multi-object and multi-character consistency, accurate text rendering within images, and flexible resolution control up to 4K. It is now integrated across Google’s AI platforms including the Gemini app, Search AI Mode, and other Gemini-powered services.
Average savings vs typical market rates
README
Overview
Nano Banana 2 is a high-performance image generation and editing model built for structured, production-grade workflows. It produces high-fidelity images with stronger prompt adherence, improved spatial reasoning, and more reliable layout control compared to earlier Nano Banana releases.
Version 2 introduces clearer text rendering, better material realism, and improved handling of complex multi-subject scenes. It is designed for practical generation tasks such as product mockups, UI concepts, structured layouts, and photorealistic compositions.
How it Works
Prompt Interpretation
Nano Banana 2 parses prompts with improved semantic and spatial reasoning. It better understands how objects relate to each other in physical space, allowing for more reliable handling of reflections, transparency, perspective, and structured compositions.
Clear, specific prompts generally produce more consistent and controllable results.
Image Generation
The model generates high-resolution images with detailed material rendering and lighting consistency. It supports both realistic and stylized outputs, maintaining composition stability even in dense scenes.
Text inside images is rendered more sharply and with stronger layout discipline than earlier versions.
Scene Logic and Real-World Grounding
Nano Banana 2 demonstrates improved real-world reasoning. Transparent objects refract correctly, reflections behave more naturally, and multi-subject interactions maintain coherent perspective. This makes it more reliable for scenes involving physical logic or structured layouts.
Key Features
- Stronger Prompt Adherence
Handles complex and structured requests with improved scene logic. - SOTA Text Rendering
Produces sharper, more usable in-image typography for posters, UI, packaging, and infographics. - Material and Lighting Realism
Improved handling of reflections, transparency, layered compositions, and surface detail. - Multi-Subject Stability
Maintains perspective and consistency in dense or interactive scenes. - High-Resolution Output
Supports outputs up to 4K resolution. - Improved Speed and Efficiency
Faster generation with improved quality-to-cost performance.
How to Use
- Provide a detailed text prompt describing the scene or layout.
- Specify any structured elements such as text content, composition, or materials.
- Run the generation.
- Refine the prompt if adjustments to layout, lighting, or structure are needed.
For structured layouts or typography-heavy scenes, explicitly define text hierarchy and placement for best results.
Example prompt:
“A modern event poster titled ‘Future Systems Summit 2026’ with clear typography hierarchy, date and location information, minimalist layout, sharp readable text, and structured composition.”
Documentation
You can find full usage details, parameters, and examples here:
https://runware.ai/docs/providers/google#nano-banana-2-gemini-31-flash-image