Google has launched Nano Banana 2, its latest image generation model. Technically called Gemini 3.1 Flash Image, the model aims to combine the image quality of the existing Nano Banana Pro with the speed of Google’s Flash model tier.
Google first released Nano Banana in August 2025. The model generated widespread interest, particularly in countries like India. A more capable version, Nano Banana Pro, followed in November 2025. Nano Banana 2 now replaces Nano Banana Pro as the default image model across Google’s Gemini app for its Fast, Thinking, and Pro modes.
What the new model can do
According to Google, Nano Banana 2 brings several improvements over the original Nano Banana:
- World knowledge and web search grounding: The model draws on Gemini’s knowledge base and real-time web search to generate more accurate depictions of real-world subjects. This also enables the creation of infographics and data visualizations.
- Text rendering and localization: The model can produce legible text within images and translate that text into other languages directly in the image.
- Subject consistency: The model can maintain the appearance of up to five characters and up to 14 objects across a single workflow, which Google says supports storytelling and storyboarding.
- Resolution and aspect ratio control: Images can be generated at resolutions from 512px to 4K in a range of aspect ratios.
Google says visual quality has also improved, with sharper details, richer textures, and more vibrant lighting compared to the original Nano Banana.
Where it is available
Nano Banana 2 is rolling out across several Google products. In the Gemini app, it becomes the new default for image generation. Users on Google AI Pro and Ultra plans can still access Nano Banana Pro for specialized tasks through the app’s menu. In Google Search, the model is available through AI Mode and Google Lens across 141 countries and in eight additional languages. The model is also available for developers via the Gemini API, AI Studio, Vertex AI, and Google’s development tool Antigravity. Google has made it the default image generation model in Flow, its video editing tool.
All images generated with Nano Banana 2 carry a SynthID watermark, Google’s system for marking AI-generated content. The images also support C2PA Content Credentials, a standard maintained by an industry group that includes Adobe, Microsoft, Google, OpenAI, and Meta. Google says its SynthID verification feature in the Gemini app has been used more than 20 million times since launching in November 2025.
Sources: Google Blog, Google Blog, TechCrunch
