Nano Banana 2

Nano Banana 2 AI Image Generator

Create high-quality images with Nano Banana 2, Google’s Gemini 3.1 Flash Image model. Generate and edit visuals from text and image references, work with multiple subjects or products, render useful text, and create 4K-capable assets with a workflow designed to balance quality, intelligence, and speed.

4K-Capable•Multiple References•Image Editing•Text Rendering•Character Consistency

Nano Banana 2 • Gemini 3.1 Flash Image

Create an Image See What It Can Do Explore original Nano Banana

NB2

4K-Capable Visual

NB2

Text + Design

NB2

References

Create with Nano Banana 2

Start from a text prompt or supported visual reference. Describe the image you want, add reference material when it improves creative control, and use the image size, aspect ratio, editing, or other options currently available in the DoMax AI generator.

Generate

Create a new image from a written idea.

Edit

Upload an existing image and describe the changes you want.

Create with References

Use supported reference images to guide characters, products, objects, composition, or other visual details.

Loading your workspace…

Why Choose Nano Banana 2?

Nano Banana 2 is designed as the general-purpose workhorse of Google’s current Nano Banana image family, combining higher-quality visual generation with Flash-class speed and broad creative control.

High-Resolution Creation

Generate at model-level resolutions ranging from lightweight previews to 2K and 4K-capable production assets.

Multiple References

Use several source images when a creative brief includes specific products, people, objects, or visual elements.

Character Consistency

Maintain recognizable people or characters more effectively across new scenes, poses, and compositions.

Reliable Text Rendering

Create posters, infographics, diagrams, menus, ads, and other visuals where readable text is an important part of the image.

Image Editing

Refine generated or uploaded imagery conversationally instead of restarting every visual from scratch.

World Knowledge

Use Gemini’s broader understanding—and model-level Search grounding where integrated—to create visuals informed by concepts and real-world information.

See Nano Banana 2 in Action

Nano Banana 2 is designed for both creative image generation and structured visual tasks where references, text, consistency, or detailed instructions matter.

NB2

4K-Capable

Create Detailed High-Resolution Visuals

Move from lightweight ideation to larger image assets when the workflow needs more detail or production resolution.

NB2

Multiple References

Combine Characters, Products and Objects

Use several reference images to define the visual elements that should appear in one new composition.

NB2

Text + Design

Create Structured Visual Content

Generate posters, infographics, diagrams, menus, advertisements, and other designs where text and image composition work together.

What Is Nano Banana 2?

Nano Banana 2 is Google’s Gemini 3.1 Flash Image model, introduced in February 2026 and later released as a stable Gemini model. It combines Gemini 3-series intelligence with fast native image generation and editing.

Model ID

gemini-3.1-flash-image

Text + Image Input

Create images using written instructions and visual source material.

Image Editing

Generate and refine images through conversational, multi-turn instructions.

High-Resolution Output

The model supports 0.5K, 1K, 2K, and 4K output sizes at the model level.

Multiple References

Gemini 3 image workflows support large reference sets, with Nano Banana 2 designed to preserve multiple objects and characters in a single workflow.

Advanced Text Rendering

Create more reliable stylized text for posters, infographics, menus, diagrams, and marketing creative.

Search Grounding

At the model level, Nano Banana 2 can use Google Web Search and Google Image Search grounding to inform image generation.

Video Context

Gemini 3.1 Flash Image supports video input at the model level, enabling video-informed image generation workflows.

Create Images up to 4K with Nano Banana 2

Nano Banana 2 supports multiple model-level resolution tiers, allowing the same image model to serve quick visual iteration and larger production assets.

0.5K

Best for

Fast previews, small concepts, thumbnails, and lightweight iteration.

Best for

Everyday image generation, web imagery, social creative, and general visual work.

Best for

Larger campaign assets, presentation graphics, detailed products, and design work.

Best for

High-resolution visual assets, large-format compositions, detailed creative, and workflows that benefit from additional image detail.

These are Gemini 3.1 Flash Image model-level capabilities. DoMax AI users should only select resolutions currently exposed by the live generator.

Create with Multiple Reference Images

Nano Banana 2 can work with a larger set of visual references than the original Nano Banana model, making it better suited to scenes with several specific characters, products, or objects.

NB2

Character A

NB2

Character B

NB2

Product

NB2

Environment

Nano Banana 2

NB2

Generated Result

Model-Level Reference Capability

up to 10 object images with high-fidelity reference handling
character resemblance for up to 4 characters in one workflow
up to approximately 14 total image references across supported Gemini 3 image workflows

Characters

Combine several recognizable people or characters into a new image.

Products

Use different product images when exact visual objects need to appear in one creative asset.

Objects

Provide source material for props, furniture, clothing, accessories, or other visual elements.

Composition

Use reference images to help communicate scene structure, framing, environment, or visual relationships.

More references do not automatically produce a better image. Give each reference a clear purpose and avoid contradictory source material.

Keep Characters and Products Consistent

Nano Banana 2 improves consistency across reference-heavy workflows, making it useful when recognizable subjects need to survive larger creative changes.

Characters

Move the same character between environments, poses, outfits, and story concepts.

People

Use reference images to maintain recognizable facial and visual traits across creative scenes.

Products

Build campaigns or lifestyle imagery around an existing product while attempting to preserve its important design details.

Objects

Reuse recognizable visual objects across compositions, ads, presentations, or design variations.

NB2

Source

NB2

Scene 1

NB2

Scene 2

NB2

Scene 3

Consistency is improved, not guaranteed. Review identity, product geometry, logos, typography, materials, and other critical details before production use.

Edit Images with Natural Language

Nano Banana 2 supports conversational image editing, allowing a visual to develop through a sequence of focused instructions.

NB2

Original

NB2

Edited

Add

Introduce new objects, subjects, text, or scene elements.

Remove

Remove unwanted people, objects, text, or visual details.

Replace

Swap backgrounds, clothing, materials, products, objects, colors, or other components.

Restyle

Change the photographic, illustrative, graphic, cinematic, or artistic direction.

Recompose

Adjust framing, positioning, scene structure, or relationships between important elements.

Refine

Continue editing the same image through additional instructions instead of restarting from the beginning.

Create Posters, Infographics and Images with Text

Text rendering is one of the areas where Nano Banana 2 expands beyond simple image generation.

Posters

Create imagery, headlines, visual hierarchy, and supporting copy in one design.

Infographics

Turn concepts, processes, comparisons, or information into structured visual communication.

Diagrams

Generate visual explanations with labels, relationships, and organized information.

Menus

Create food, service, product, or promotional layouts where imagery and readable text work together.

Ad Creative

Combine product visuals, campaign messaging, and graphic composition.

Multilingual Graphics

Create visual concepts containing text across supported languages and localized design workflows.

Text rendering is improved but not infallible. Review spelling, numbers, brand names, factual data, hierarchy, and typography before publishing.

Model-level capability

Create Images with Search-Grounded Information

At the Gemini model level, Nano Banana 2 can use Google Search to bring current web information into an image-generation workflow. Gemini 3.1 Flash Image can also use Google Image Search as visual grounding.

Current Information

Create charts, explanatory visuals, or graphics informed by recent information.

Real-World Subjects

Use search grounding to improve factual context around places, objects, species, products, or other real-world topics.

Visual Research

Google Image Search grounding can provide visual context for image generation at the underlying model level.

Information Graphics

Combine current information with Nano Banana 2’s text rendering and layout capabilities.

Model-level capability

Create Images from Video Context

Gemini 3.1 Flash Image supports video input at the model level, allowing a video to provide multimodal context for generating a new still image.

Thumbnails

Use video context to create thumbnail concepts related to the source footage.

Posters

Turn video context into a cinematic poster or promotional still.

Summary Graphics

Use video content as context for a visual summary, educational graphic, or supporting creative asset.

Nano Banana 2 Aspect Ratios and Resolution

Nano Banana 2 supports both common image formats and unusually wide or tall compositions at the model level.

Resolutions

0.5K1K2K4K

Model-Level Aspect Ratios

1:1

Square

2:3

Portrait

3:2

Landscape

3:4

Portrait

4:3

Landscape

4:5

Portrait

5:4

Landscape

9:16

Vertical

16:9

Wide

21:9

Ultra-wide

1:4

Tall

Extreme

4:1

Wide

Extreme

1:8

Ultra-tall

Extreme

8:1

Ultra-wide

Extreme

These describe the underlying Gemini 3.1 Flash Image model. The DoMax AI generator is the source of truth for the resolutions and aspect ratios currently exposed on this page.

Best Use Cases for Nano Banana 2

NB2

Product & Advertising

Create studio images, lifestyle scenes, campaign concepts, product compositions, and promotional visuals using one or more product references.

NB2

Characters & Storytelling

Develop recurring characters across scenes, outfits, poses, environments, and narrative concepts.

NB2

Posters & Infographics

Combine imagery, readable text, structured layouts, and visual information in one generated asset.

NB2

Image Editing

Refine uploaded or generated imagery through natural-language changes and multi-turn creative iteration.

NB2

Marketing & Social Creative

Create campaign visuals, social graphics, announcements, covers, ads, and visual variations in different formats.

NB2

High-Resolution Assets

Use 2K or 4K-capable model output when the creative workflow benefits from additional image detail and larger dimensions.

How to Use Nano Banana 2 on DoMax AI

Nano Banana 2 AI image generator on DoMax AI

Use the live DoMax AI generator to review the controls currently available.

Choose Nano Banana 2

Open the Gemini 3.1 Flash Image workflow when you need its generation, editing, reference, text-rendering, or high-resolution capabilities.

Choose how to start

Write a prompt or upload supported source images for editing and reference-based creation.

Describe the final image

Define the subject, environment, composition, style, lighting, text, important details, and creative purpose.

Add references and settings

Use reference images intentionally and select the resolution, aspect ratio, or other options currently available in the DoMax AI generator.

Generate and refine

Review composition, subject consistency, text, product details, identity, and visual hierarchy. Continue with focused edits when most of the image already works.

Tips for Better Nano Banana 2 Results

Assign Every Reference a Role

Explain which image defines the character, product, object, environment, or other creative detail.

Describe the Final Composition

State how subjects should be arranged, framed, sized, and positioned.

Use Exact Text

For posters and graphics, provide important copy exactly as it should appear.

Iterate Conversationally

When most of an image works, make focused changes instead of rewriting the entire creative brief.

Choose Resolution for the Job

Use lightweight resolutions for rapid exploration and larger outputs when additional detail is actually useful.

Review Factual Visuals

Search grounding can improve context, but important numbers, labels, dates, maps, charts, and factual statements should still be checked.

Check Identity and Products

Inspect faces, products, logos, materials, text, hands, geometry, and other production-critical details before publishing.

Nano Banana 2 Limitations

Character and product consistency is improved but not guaranteed across every complex transformation.
Multiple references can conflict when source images contain inconsistent identity, style, lighting, perspective, or product details.
Text rendering is more reliable, but exact spelling, numbers, typography, alignment, and dense layouts can still require iteration.
The model may not always produce the exact number of image outputs requested.
Complex layouts and precise object placement may vary from the requested dimensions or spatial relationships.
Search-grounded generation can provide useful context, but important real-world information should still be verified.
Google Search grounding for Gemini 3.1 Flash Image currently has limitations around real-world images of people from web search.
Video input is supported at the underlying model level, but it should not be presented as available on DoMax AI unless the current integration exposes it.
Model support for 4K does not mean every DoMax generation should use or expose 4K.
Reference limits, resolution, aspect ratio, uploads, credits, pricing, and other generation settings must follow the live DoMax AI interface.

How Nano Banana 2 Fits the Nano Banana Family

Nano Banana 2

General-purpose workhorse

high-quality everyday generation
4K-capable output
multiple references
character consistency
text rendering
editing
broad production workflows

Original Nano Banana

Earlier fast image workflow

straightforward generation
natural-language editing
simpler reference workflows
approximately 1024px-class output

Choose a Nano Banana model based on the amount of reference control, image resolution, speed, efficiency, and production precision your workflow actually needs.

Explore original Nano Banana

More Nano Banana 2 Examples

Explore Nano Banana 2 across high-resolution imagery, editing, characters, product creative, typography, and structured visual design.

NB2

Product

NB2

Character

NB2

Editing

NB2

Poster

NB2

Infographic

Nano Banana 2 FAQ

What is Nano Banana 2?

Nano Banana 2 is Google’s Gemini 3.1 Flash Image model. It is a general-purpose AI image generation and editing model designed to combine fast generation with high-quality output, multiple references, character consistency, reliable text rendering, and resolutions up to 4K at the model level.

Is Nano Banana 2 the same as Gemini 3.1 Flash Image?

Yes. Nano Banana 2 is the product name used for Gemini 3.1 Flash Image. The model ID is gemini-3.1-flash-image.

Can Nano Banana 2 generate 4K images?

Yes. Gemini 3.1 Flash Image supports 0.5K, 1K, 2K, and 4K image sizes at the model level. Check the live DoMax AI generator to confirm which resolutions are currently available here.

Can Nano Banana 2 edit existing images?

Yes. Nano Banana 2 supports image editing and conversational multi-turn refinement. You can use source images together with natural-language instructions to change or develop a visual.

How many reference images can Nano Banana 2 use?

At the Gemini model level, Nano Banana 2 can work with large reference sets, including high-fidelity handling for up to 10 object images and character resemblance for up to 4 characters in a workflow. The exact upload limit on DoMax AI should follow the live generator.

Is Nano Banana 2 good for character consistency?

Yes. Character consistency is one of the model’s key reference workflows. Gemini 3.1 Flash Image can preserve the resemblance of multiple characters in a single workflow, although exact consistency is not guaranteed.

Is Nano Banana 2 good at generating text inside images?

Nano Banana 2 includes improved text rendering for structured visuals such as posters, infographics, diagrams, menus, and marketing assets. Important spelling, numbers, typography, and layouts should still be reviewed.

Can Nano Banana 2 use Google Search?

At the underlying Gemini model level, Nano Banana 2 supports grounding with Google Search and can also use Google Image Search as visual context. Whether Search grounding is available on DoMax AI depends on the current integration.

Can Nano Banana 2 create an image from a video?

Gemini 3.1 Flash Image supports video input at the model level, enabling workflows such as video-informed thumbnails, posters, and visual summaries. Check the DoMax AI generator to see whether video input is currently exposed here.

What is the difference between Nano Banana and Nano Banana 2?

The original Nano Banana uses Gemini 2.5 Flash Image and is optimized around fast 1024px-class image generation and editing. Nano Banana 2 uses Gemini 3.1 Flash Image and adds higher-resolution output up to 4K, larger reference workflows, improved consistency, stronger text rendering, broader aspect ratios, and additional model-level capabilities.

Explore More Nano Banana Models

Original workflow

Original Nano Banana

For fast image generation, natural-language editing, subject consistency, and simpler reference workflows.

Explore Nano Banana

Create with Nano Banana 2

Generate and edit high-quality images with multiple references, character consistency, text rendering, broad aspect ratios, and up to 4K model-level output.

Create an Image Explore Original Nano Banana