Google Unveils Nano Banana 2 With Faster AI Imaging & Better Text

In August of last year, Google debuted the Nano Banana AI image generator. At introduction, the business claimed that the new tool, also called the Gemini 2.5 Flash Image AI model, offers increased character consistency capabilities, element-based image modification, and faster image generation. Later, in November 2025, the search giant began releasing the Gemini 3 Pro-powered Nano Banana Pro, dubbed the state-of-the-art (SOTA) picture generation and editing model, which can produce photographs with up to 4K resolution. Google, a tech company based in Mountain View, has just released Nano Banana 2, a new AI picture-generating tool that improves the quality of the Pro model at a quicker rate.

The Nano Banana 2 (also called Gemini 3.1 Flash Image) was released by Google. This model gives the quicker, more affordable Flash-tier model “Pro” capabilities.

Google revealed that Nano Banana 2 has begun to roll out in a blog post on Thursday. Through a number of Google products, users may now test the company’s most recent Gemini 3.1 Flash Image AI model, also known as Nano Banana 2. For the Gemini 3 series Fast, Thinking, and Pro AI models in the Gemini app, Nano Banana 2 takes the place of Nano Banana Pro. Users of Google AI Pro and Ultra will still be able to utilise it for “specialised tasks”, though, by using the app to regenerate photos.

The Nano Banana 2 AI picture editing and creation tool is also accessible through the Google app, mobile browsers, and PC browsers in 141 additional countries, including India, and in eight additional languages in Search in AI Mode and Google Lens. A preview of Nano Banana 2 is also available in Vertex AI, Gemini API, and AI Studio.

It will be accessible to all Flow users for zero credits and is now the default picture production model in Google Labs’ high-fidelity video creation tool, Flow. Finally, Google Ads also offers Nano Banana 2. According to the IT giant, it can offer suggestions for the advertisements and assist customers in creating campaigns and banners.

Regarding its features and enhancements, Google stated that Nano Banana 2 makes use of “Gemini’s advanced knowledge base” and can make use of real-time data and pictures from web searches. This enables improved subject-wise rendering capabilities to be provided more quickly by the AI image editing and generating tool. 

Furthermore, it makes it possible for Nano Banana 2 to produce more detailed infographics, graphs, and data visualisations.

Furthermore, Nano Banana 2 can produce “legible” and “precise” text for greeting cards, marketing campaigns, and other materials. Additionally, it enables users to contextualise and translate words within an image. According to Google, Nano Banana 2 can also quickly produce high-quality “photorealistic imagery”. In a single workflow, it may preserve visual fidelity of up to 14 objects while retaining subject consistency of up to five characters.

Additionally, Nano Banana 2 has enhanced “instruction following” capabilities. It can comprehend subtleties in user prompts and adhere to more intricate user demands. Google noted that Nano Banana 2, like the Nano Banana Pro tool, can produce photographs with up to 4K quality and incorporate better textures, vivid lighting, and sharper details.

Some important enhancements that comes with the Nano Banana 2 are,

• Performance and Speed: This model combines the speed of the Flash model with the logic of the Nano Banana Pro. It has been incorporated into Google Ads and is currently the default picture creation model in Google Labs’ Flow.
• Improved Text Rendering: Nano Banana 2 can produce clear, accurate text right inside pictures. It facilitates in-image localisation and translation as well as marketing materials.
• Real-World Grounding: To more accurately generate subjects, infographics, and data visualisations, the model makes use of Gemini’s knowledge base and real-time online search.
• Subject Consistency: It is appropriate for storyboarding since it preserves visual integrity for up to 14 objects and character likeness for up to five characters in a single workflow.
• Production Quality: Upscaling from 512px to 4K is supported. A variety of aspect ratios are also supported (16:9, 9:16, 2:1).
• Content Transparency: To guarantee transparency, each image has C2PA Content Credentials and SynthID watermarking.

The Google ecosystem offers access to Nano Banana 2. This comprises Vertex AI, Google AI Studio, Google Search (AI Mode and Lens), and the Gemini app.