Google Sets 'Nano Banana 2' as Default Image Tool

Officially dubbed Gemini 3.1 Flash Image, the new model merges the rapid generation speeds of Google's Gemini Flash architecture with the advanced features previously found in the higher-tier Nano Banana Pro. This integration aims to close the gap between speed and visual fidelity, making high-quality image generation a default, low-latency experience. A key technical improvement is enhanced subject consistency, a significant challenge in generative AI. Nano Banana 2 can maintain the appearance of up to five distinct characters and the fidelity of 14 objects within a single narrative workflow, allowing for more coherent storyboarding and sequential image creation. The model also demonstrates superior instruction following and leverages Google Search for real-time world knowledge, allowing it to generate more accurate infographics and diagrams grounded in current information. For developers, output resolutions are now more flexible, ranging from 512px up to 4K, with expanded support for various aspect ratios. Deployment is happening at scale across Google's ecosystem, making Nano Banana 2 the default generator in the Gemini app, Google Lens, and AI Mode in Search across 141 countries. It's also integrated into developer- and enterprise-focused platforms, including AI Studio, Google Cloud's Vertex AI, and the API for Google Ads. For developers building on the platform, Nano Banana 2 is accessible via the Gemini API and Vertex AI. While this new model is the default for all users, Google AI Pro and Ultra subscribers retain the option to regenerate images using the specialized Nano Banana Pro model for specific high-fidelity tasks. To address content authenticity, outputs from Nano Banana 2 are digitally watermarked using Google's SynthID technology. This invisible watermark is designed to help identify AI-generated content, a critical MLOps and safety consideration as generative models become more widespread.