Gemini Nano Banana 2.1

Nano Banana 2.1 is Google's latest high-efficiency model for image generation and conversational editing. Use it to create 1K, 2K, or 4K images, combine up to 14 references, and refine visuals across multiple turns. GPTProto offers pay-as-you-go access at 40% below Google's list price through one API key.

Modalidades

Entrada

Prompt*

Aspect_Ratio

Size

Output_Formart

Enable_Base64_OutputIf enabled, the output will be encoded into a BASE64 string instead of a URL. This property is only available through the API.

Enable_Sync_ModeIf set to true, the function will wait for the result to be generated and uploaded before returning the response. It allows you to get the result directly in the response. This property is only available through the API.

Saída

Exemplos de uso da API

curl --request POST "https://gptproto.com/api/v3/google/gemini-nano-banana-2.1/text-to-image" \
    --header "Authorization: Bearer $GPTPROTO_API_KEY" \
    --header "Content-Type: application/json" \
    --data '{
        "prompt": "A tiny origami fox sailing a teacup across a moonlit puddle",
        "aspect_ratio": "16:9",
        "size": "1k",
        "output_formart": "png",
        "enable_base64_output": false,
        "enable_sync_mode": false
    }'

Preços Gemini Nano Banana 2.1

Comece pelo custo de uma única amostra e escolha um orçamento de teste. As tarifas da GPTProto são 40% abaixo do preço de tabela.

Tamanho

Cenário	Lista de Google	OpenRouter	GPTProto	Economia / mês
Testes rápidos e protótipos100 imagens / mês	$3.36	$3.54	$2.02	−$1.34≈ $16.13 / ano
Fotos de produtos e e-commerce1,000 imagens / mês	$33.60	$35.45	$20.16	−$13.44≈ $161.28 / ano
Criativos de anúncios / marketing2,500 imagens / mês	$84.00	$88.62	$50.40	−$33.60≈ $403.20 / ano
Conteúdo de redes sociais10,000 imagens / mês	$336.00	$354.48	$201.60	−$134.40≈ $1612.80 / ano

Custo mensal por fonte

OpenRouter

$0.00

Google

$0.00

GPTProto

$0.00

Economize $0.00 / mês

Economia anual ≈ $0.00 · pague pelo uso

Nano Banana 2.1 API - 4K Image Generation at 40% Off

Google's gemini-nano-banana-2.1 improves visual quality, prompt adherence, text rendering, and consistency across edits while retaining Flash-level efficiency. It is designed as the everyday production model below the more expensive Nano Banana Pro tier.

Up to 14 Reference Images

Combine up to 10 object references and four character references in one workflow. This is useful for product fidelity, recurring characters, campaign variations, and multi-subject compositions.

1K to 4K and Extreme Panoramas

Choose 1K, 2K, or 4K output. Nano Banana 2.1 also improves wide and panoramic ratios such as 1:4, 4:1, 1:8, and 8:1 at higher resolutions.

Cleaner Text and Structured Layouts

Generate posters, menus, product graphics, and infographics with improved text rendering and layout accuracy. Review small print, prices, and legal copy before publishing.

40% Below Google List Price

GPTProto's listed rates start at about $0.0202 for a 1K image, with 2K and 4K options available. Use the live calculator to confirm the final cost for your request.

What Is the Nano Banana 2.1 API

Nano Banana 2.1 is Google's stable, high-efficiency image generation and editing model released on October 6, 2026. It updates Nano Banana 2, whose official model name is Gemini 3.1 Flash Image. The new version improves realism, instruction following, typography, infographic layouts, and character consistency during multi-turn edits.

The model accepts text, image, video, and PDF inputs and returns image or image-plus-text output. It can generate a new visual, edit an existing one, combine multiple references, or use a previous interaction as context for the next revision. Its configurable thinking levels - minimal, medium, and high - let developers balance latency against the planning needed for a complex composition.

Google also documents Web Search and Image Search grounding for Nano Banana 2.1. This can help with visuals that depend on current or factual context, but generated labels, facts, charts, and numbers still require review. When search grounding is used in a user-facing product, Google's required search suggestions and attributions must be displayed.

Is gemini-nano-banana-2.1 the Same Model

Yes. gemini-nano-banana-2.1 is Google's official model ID for Nano Banana 2.1. Do not confuse it with Gemini 3 Pro Image, also called Nano Banana Pro. Nano Banana 2.1 is the efficient production model; Nano Banana Pro is the premium option for the most complex brand, localization, and precision-design work.

Nano Banana 2.1 Specifications

Specification	Nano Banana 2.1
Provider	Google
Release date	October 6, 2026
Official model ID	`gemini-nano-banana-2.1`
Main tasks	Text-to-image, image editing, multi-turn editing, multi-image fusion, video-to-image
Supported inputs	Text, image, video, and PDF
Outputs	Image and text
Output resolutions	1K, 2K, and 4K; 1K by default
Reference images	Up to 14 total: up to 10 objects and four characters
Thinking levels	`minimal`, `medium` by default, and `high`
Search grounding	Google Web Search and Image Search
Token limits	131,072 input; 32,768 output
Provenance	Invisible SynthID is included in generated images

Route note: The table describes the underlying Google model. Check the fields shown in GPTProto's Playground and API Usage schema before making search grounding, video or PDF input, thinking levels, or all 14 references a required part of your production workflow.

Nano Banana 2.1 vs Nano Banana Pro vs GPT Image 2.5

These models overlap, but they are not interchangeable. Nano Banana 2.1 is the cost-efficient choice for frequent generation and iterative editing. Nano Banana Pro targets demanding design work. GPT Image 2.5 is split into Flare for faster everyday generation and Sunburst for tighter control over detailed edits.

Model	Best Fit	Resolution and Editing	GPTProto Billing
Nano Banana 2.1	High-volume generation, infographics, product variations, recurring characters, and conversational edits	1K, 2K, and 4K; up to 14 references	Per-image rates listed 40% below Google
Nano Banana Pro	Complex layouts, localization, brand consistency, and quality-first design work	1K, 2K, and 4K; premium precision controls	From $0.0804 per image, 40% below Google
GPT Image 2.5 Flare	Fast generation and precise everyday edits in an OpenAI workflow	Text and image inputs with selectable size and quality	Token-based rates listed 5% below OpenAI
GPT Image 2.5 Sunburst	Polished campaign assets and edits where precision matters more than speed	Text and image inputs; longer generation for quality-first work	Token-based rates listed 5% below OpenAI

Choose Nano Banana 2.1 when predictable per-image cost, Google Search grounding, 4K output, or a large reference set matters. Choose Nano Banana Pro when the brief depends on premium brand or localization control. Choose GPT Image 2.5 when your application already uses OpenAI's image workflow or needs Flare and Sunburst as separate speed and precision options.

How to Write a Nano Banana 2.1 Prompt

A useful Nano Banana 2.1 prompt states the asset type, subject, layout, exact text, visual style, lighting, aspect ratio, and details that must not change. For editing, identify the requested change first and list the protected elements second. Make one category of change per turn when subject identity or product geometry must remain stable.

Prompt example

Create a 4:5 launch poster for a citrus sparkling-water can. Place one cold silver can in the lower center on pale stone, with realistic condensation and a soft morning shadow. Add the exact headline "BRIGHT, NOT SWEET" in large navy type at the upper left. Keep the label fully readable. Use an ivory, navy, and citrus-yellow palette. Return a polished 2K image with no extra logos, badges, or text.

For a revision, continue with a focused instruction such as: "Keep the can, label, headline, layout, and lighting unchanged. Replace only the pale-stone surface with brushed stainless steel." This gives the multi-turn editor a clear boundary.

FAQ

Is Nano Banana 2.1 an Official Google Model

Yes. Google released the stable `gemini-nano-banana-2.1` model on October 6, 2026. It replaces Nano Banana 2 as Google's recommended high-efficiency image workhorse for new projects.

How Do I Access the Nano Banana 2.1 API on GPTProto

Create a GPTProto API key, select Nano Banana 2.1 in the Playground or API Usage section, and use the endpoint and request fields shown on this page. The same GPTProto key and balance can also be used with other supported image, video, and language models.

How Much Does the Nano Banana 2.1 API Cost

Google's standard image-output list rates are $0.0336 for 1K, $0.0504 for 2K, and $0.113 for 4K. At GPTProto's listed 40% discount, the corresponding base rates are approximately $0.0202, $0.0302, and $0.0678. Check the live pricing module for the exact request total.

Can the Nano Banana 2.1 API Generate 4K Images

Yes. The model supports native 1K, 2K, and 4K output, with 1K as the default. Select 4K when the final asset needs more detail; use 1K or 2K for drafts, previews, and higher-volume jobs.

Does the Nano Banana 2.1 API Work Without a Watermark

Generated images do not need a visible corner logo, but they are not watermark-free. Google states that all Nano Banana outputs include invisible SynthID provenance. The API should not be marketed as removing that watermark.

Nano Banana Pro vs Nano Banana 2.1 Which Should I Use

Use Nano Banana 2.1 for an affordable balance of quality, speed, 4K output, and iterative editing. Use Nano Banana Pro for the most demanding layouts, localization, brand consistency, or precision creative control. Test the same brief on both when the final asset is business-critical.

Nano Banana 2.1 vs GPT Image 2.5 Which Is Better

Nano Banana 2.1 is a strong default for per-image pricing, Google-grounded workflows, large reference sets, and 4K production. GPT Image 2.5 Flare is designed for faster everyday work, while Sunburst prioritizes detailed, controlled edits. The better choice depends on your prompts, edit pattern, latency target, and billing preference.

Does Nano Banana 2.1 Support Image Editing and Multiple References

Yes. The underlying model supports conversational image editing and up to 14 reference images, including up to 10 object references and four character references. Confirm the maximum accepted by the specific GPTProto route before sending a production batch.