Google DeepMind releases Nano Banana Pro, advancing AI image generation with studio-quality output

Google DeepMind has launched Nano Banana Pro (Gemini 3 Pro Image), a next-generation image generation and editing model built on Gemini 3 Pro. The model is rolling out in paid preview to developers through the Gemini API, Google AI Studio, and Vertex AI for enterprise users.

Key capabilities#

The model offers several significant improvements over its predecessor, Gemini 2.5 Flash Image:

• High-fidelity output with support for 2K and 4K resolution, giving developers granular control over lighting, camera focus, color grading, and composition
• Advanced text rendering that produces clear, accurate text integrated into images—a major leap for functional asset generation
• Localization support enabling semantic understanding of images for language changes in menus, signs, and documents
• Grounding with Google Search, connecting the model to real-time web content for factual, data-driven outputs

Expanding the ecosystem#

Nano Banana Pro is gaining traction across the developer landscape. Google is integrating it into Antigravity, its new agentic development platform, where coding agents can generate UI mockups and visual assets. Creative platforms including Adobe and Figma are also incorporating the model.

The release includes demo applications showcasing capabilities like comic book generation, infographic creation, and product mockup design.

AI provenance and transparency#

All images created or edited with Gemini 3 Pro Image include SynthID digital watermarks to denote their AI-generated or edited origin, addressing transparency concerns in synthetic media.

Developers can choose between Gemini 2.5 Flash Image for faster, lower-cost generation or Gemini 3 Pro Image for higher-quality outputs. More details are available in the official documentation.