Google Now it can truly be said that there’s no letup for competitors, with updates coming every few weeks for their lineup AI Products. Then Gemini 3.1 Pro After that, earlier once again announced the launch of the brand new ‘Nano Banana 2 The image generation model (Gemini 3.1 Flash Image) not only has improved capabilities across multiple areas, but the generation speed has also become significantly faster. More importantly, now… Gemini Even with the “Fast” model, you can now generate images with Traditional Chinese characters—no Pro subscription needed, which is great news for free users (the daily generation limit has increased but is still in place).
Google launches Nano Banana 2, its next-generation image generation model, now available to free users as well
While Nano Banana was fast at generation, it had limited world knowledge and struggled with complex details, and couldn’t generate images with traditional Chinese characters, making it only suitable for simple edits and basic consistency. Nano Banana Pro offers better quality, reasoning, and text rendering capabilities, which is why many people use the Pro version for image generation with Gemini.
The newly launched Nano Banana 2 model has officially merged the advantages of these two older models. In short, it not only has Pro-level features (such as real-time network knowledge, precise instructions, text capabilities, and topic consistency), but also Flash-level fast generation speed, significantly reducing wait times and allowing users to quickly generate multiple versions.
In Gemini, generating images using any of the modes—Flash, Think, or Pro—will activate the Nano Banana 2 model:
Even in quick mode, it can still generate perfectly rendered traditional Chinese character images:
Here are the main features of Nano Banana 2:
- Experience the high-end professional capabilities of Nano Banana Pro combined with the ultimate generation speed of Gemini Flash.
- Possessing powerful world knowledge capabilities, enabling real-time extraction of accurate information and images from Google web searches to ensure the generated content is more precise and贴近 reality.
- Excels at maintaining consistency in character and object appearance, keeping the same person, animal, or object looking highly similar across multiple images, sequential stories, or different scenes with consistent facial features, clothing, poses, and details. The official states it can handle up to 5 character appearances and 14 object details consistently.
- Precisely understand complex natural language instructions, so users don’t need technical jargon—they can simply describe the desired changes in everyday language, and the model will correctly identify what to modify and what to preserve. Editing images also doesn’t require drawing masks; you can direct changes with text alone.
- Better generation quality with support for up to 4K resolution, flexible aspect ratio adjustment, and enhanced texture details and sharpness.
For Gemini users, you might be wondering: since you can now generate images with Nano Banana 2 just by selecting “Quick,” what’s the difference if you switch to Pro?
When using Pro to generate images, there’s no difference the first time—it uses the Nano Banana 2 model. But if you’re not satisfied with the image, Google AI Pro and Ultra subscribers will switch to the Nano Banana Pro model when using the “Regenerate” feature.
According to Google, the differences are:
- Nano Banana Pro: High-Fidelity and Maximum Fact Accuracy Tasks
- Nano Banana 2: Fast Generation, Precise Instruction Comprehension, Integrated Image Search
Google also provides a few image examples generated with Nano Banana 2.
This image demonstrates the consistency of Nano Banana 2, with the same character appearing across different images.
Putting multiple characters in the same image:
I used Google’s “Water Cycle” prompt to generate images, and the quality is really quite good, with both the Chinese and English text being completely correct:
Nano Banana 2 is now officially available on the following platforms:
- Gemini application
- Google Search AI Mode, Lens, Google Ads, and Google Flow Creative Studio
- Developer tools such as Google AI Studio and the Gemini API
Free users can access it directly in the Gemini App and Google Search without an additional subscription, but there are daily generation limits (depending on demand). Paid users such as Google AI Pro or Ultra enjoy higher quotas, unlimited usage, and additional perks like retaining access to the original Pro model.
Here are the current Nano Banana 2 quota limits: Free up to 20 per day (Pro models unavailable), AI Plus up to 50 per day, AI Pro up to 100 per day, AI Ultra up to 1,000 per day. Google has also noted that if usage is too high, the limits may be subject to frequent adjustments:
Source: KOCPC Chinese