Nano Banana 2 (Gemini 3.1 Flash Image) Released with Enhanced Image Generation and Editing
Key Points
- High-fidelity image generation with improved world knowledge integration
- Advanced text rendering and multi-language localization support
- New aspect ratios and configurable thinking levels for better control
Summary
Google DeepMind has released Nano Banana 2 (Gemini 3.1 Flash Image), a new image generation and editing model that delivers high-fidelity visuals with improved performance and advanced features. The model is available through the Gemini API in Google AI Studio and requires a paid API key.
Key Points
- Enhanced World Knowledge: Leverages Gemini's broad knowledge base to create detailed visuals using web search references for more realistic image generation
- Advanced Text Rendering: Improved text accuracy and crispness with support for in-image localization across multiple languages
- Expanded Creative Controls:
- New aspect ratios (4:1, 1:4, 8:1, 1:8) alongside existing options
- New 512px resolution tier for optimized efficiency and reduced latency
- Configurable thinking levels (Minimal vs. High/Dynamic) for better prompt adherence
- Improved instruction following for complex, multi-layered prompts
- Production Ready: Available via Gemini API, Google AI Studio, Vertex AI, Google Antigravity, and Firebase
- Demo Applications: Includes example apps like "Window Seat", "Global Ad Localizer", and "Pet Passport" showcasing the model's capabilities