Google has officially launched the latest generation image generation model, Nano Banana2 (technical name: Gemini3.1Flash Image), in the Gemini App. This update perfectly combines the high-end Pro-level image capabilities with the Flash-level fast response, allowing more users to easily enjoy a high-quality and fast AI image creation experience.
Model directly integrated into Gemini App, more convenient operation
Users just need to select the "Image Generation" feature in the Gemini App to directly access Nano Banana2. The model name is clearly displayed as "Nano Banana2," marking this generation's official replacement of the previous version, becoming the default image generation engine. Whether it's Fast, Thinking, or Pro mode, Nano Banana2 will be used first for creation, greatly simplifying the user's steps to switch models.
Core upgrade: Default 2K resolution, more detailed clarity
Compared to the previous generation, Nano Banana Pro, the biggest highlight of Nano Banana2 is the default generation resolution increased to 2K, significantly improving image details and clarity. It also supports up to 4K upscaling, allowing users to easily output images of production-grade specifications, suitable for marketing posters, social media materials, or creative designs. The generation speed inherits the lightning-fast characteristics of Gemini Flash, making iterative creation almost wait-free.
Significant improvement in text rendering capability
Many users have reported that Nano Banana2 has made significant progress in text rendering. Whether it's complex Chinese-English mixed layout, font styles, translation localization, or signs, posters, and text embedded in images, it can present them more legible and accurate, avoiding common issues such as blurriness, distortion, or typos. This improvement is especially suitable for creating greeting cards, advertising mockups, product labels, and other scenarios requiring precise text.
More professional-level features are popularized
Nano Banana2 not only is fast but also inherits the advanced capabilities of the Pro level, including:
- Generating more accurate specific objects, landmarks, scenes using Gemini's real-time world knowledge and web search;
- Supporting subject consistency, coherent generation of multiple characters/objects (up to 5 characters + 14 objects);
- Flexible adjustment of aspect ratio, dynamic editing, and precise instruction following;
- Built-in SynthID watermark and content credentials, enhancing the transparency and traceability of AI-generated content.
How to experience immediately
Open the Gemini App → click on 🍌 "Create Image" or directly enter generation instructions in the conversation → the model automatically loads Nano Banana2. Free users can use it (with generation limits), and paid subscription users enjoy higher quotas and priority access.
AIbase review
