Gemini 3 Pro Image — Google's premium image model, one API call away
Gemini 3 Pro Image is Google's model for professional image work: 1K to 4K output, built-in thinking and up to 14 references. Send the generateContent body you already know. From $0.01875 per image.
- Text to image
- Image editing
- 1K to 4K
- Built-in thinking
- Up to 14 references
- generateContent body
Try Gemini 3 Pro Image before you write any code
Run Gemini 3 Pro Image with the exact JSON your code will send. Every generateContent field the endpoint takes is here.
gemini-3-pro-image-official$90.00 / 1M image output tokens
Describe an image, add references if you have them, and submit once. The result appears when the task completes.
What to create or change. Subject, setting, lighting and style.
Reference images the model should follow. Files are uploaded and sent as URLs.
Up to 14, sent as fileData URLs.
Default matches your first reference, otherwise 1:1.
Output resolution. 1K is the default.
Image only, or image with a short text reply.
Standing rules for the model, such as a house style.
Greyed-out options relax content filtering. They are available through the API only, not in the playground.
0–2. Leave empty for the model default.
0–1. Leave empty for the model default.
Whole number, 1 or more.
Whole number. Reuse it to reproduce a result more closely.
Caps text and thinking output.
One per line, up to 5.
Only 1: one request returns one image.
How many tokens each reference uses. Lower is cheaper.
Return the model's thought summary with the image.
IMAGE_JPEG is the one documented value.
Applies only when aspect ratio above is Default; when both are set, aspect ratio wins.
Applies only when image size above is Default; when both are set, image size wins.
Estimated at current rates from the image tokens of the selected size plus a measured sample of text and thinking tokens. Reference inputs add a little on Official.
Try one of these prompts
Sample images generated with Gemini 3 Pro Image through this API
Gemini 3 Pro Image pricing
Two ways to pay for Gemini 3 Pro Image. Standard is one flat price per image at any size. Official bills the tokens each render reports, where 1K and 2K cost the same and 4K costs more.
- Official
- $0.10304/image
- Standard
- $0.01875/image
- Official
- 194images
- Standard
- 1,066images
Both prices come from the rates that bill your account. Failed requests are not charged.
Official estimates use the image tokens Google publishes for each size, plus a measured sample of 246 text and thinking tokens and 14 prompt tokens.
Gemini 3 Pro Image cost by image size
Google publishes the image tokens for each Gemini 3 Pro Image size: 1K and 2K use 1,120, 4K uses 2,000. Official prices below add a measured thinking sample.
| Category | gemini-3-pro-image-officialPrice · 1 image | gemini-3-pro-imagePrice · 1 image |
|---|---|---|
| 1K | $0.103035 | $0.01875 |
| 2K | $0.103035 | $0.01875 |
| 4K | $0.182235 | $0.01875 |
Standard is one flat price per image at every size. Official follows the image tokens, so larger sizes cost more.
Price your own Gemini 3 Pro Image request
Pick a size and watch the figure. It is computed from the live rate and the image tokens Google publishes for each size, so it moves when the price does.
A pre-submit estimate. Standard bills each delivered image; Official bills image output, text and thinking output, and input at their own rates.
gemini-3-pro-image-official: Image output $90.00 / 1M · Text and thinking output $9.00 / 1M · Input $1.50 / 1M
What is Gemini 3 Pro Image?
Gemini 3 Pro Image is Google's image generation and editing model, sold under the name Nano Banana Pro. Google calls it the premium choice for the most complex visual tasks and designed it for professional asset production.
On SeedRouter you call Gemini 3 Pro Image with Google's own generateContent request body plus a model field. Submit once, keep the task ID, and read a hosted image URL when the task completes.
Model ID: gemini-3-pro-image · Token billing ID: gemini-3-pro-image-official · Input limit: 65,536 tokens.
Two ways to use Gemini 3 Pro Image
Call the API from your backend, or hand the job to a coding agent.
Call the API
Send the Gemini 3 Pro Image request from your server, keep the task ID, and read the finished image from the task.
- 1Create an API key
- 2POST the request body
- 3Poll the task until it completes
Let an agent do it
Copy a ready prompt into Claude Code, Codex or Cursor. It asks before spending and downloads the result.
- 1Copy the prompt
- 2Fill in what you want
- 3Approve the request
Why developers pick Gemini 3 Pro Image
The Gemini model to use when a render has to get complex instructions right.
Google's request body
Gemini 3 Pro Image takes the same contents and generationConfig you would send to Google. Add a model field and you are done.
Thinks before it draws
Gemini 3 Pro Image reasons about layout first. Ask for thought summaries with includeThoughts when you want to see why.
Legible text
Google lists legible, stylized text for infographics, menus and diagrams among the strengths of its Gemini 3 image models.
4K output
Gemini 3 Pro Image renders 1K, 2K and 4K. On token billing, 2K costs the same as 1K.
14 references
Combine up to 14 reference images in one Gemini 3 Pro Image request, passed as fileData URLs.
Task-based
Long Gemini 3 Pro Image renders never hold your connection open. You poll a task and get a URL.
What teams build with Gemini 3 Pro Image
Infographics and diagrams
Gemini 3 Pro Image turns notes into labeled explainers.
Brand-consistent sets
Google names accurate brand consistency among the strengths of Gemini 3 Pro Image, useful for campaign sets.
Print artwork
Posters and packaging at 4K from Gemini 3 Pro Image.
Localized assets
Google highlights advanced localization, so one Gemini 3 Pro Image design can carry text in several languages.
Product scenes
Place a product from a reference photo into new Gemini 3 Pro Image scenes.
Precise edits
Change one element of an image with Gemini 3 Pro Image and keep the rest.
Start with Gemini 3 Pro Image in four steps
- 01
Create a key
Sign in, top up any amount, and create an API key. Credits never expire.
- 02
Send one request
POST the Gemini 3 Pro Image body to /v1/images/generations with your prompt and any reference URLs.
- 03
Poll the task
Keep the task ID and check it every few seconds until the status is completed.
- 04
Download the image
Save the hosted image URL from the task output to your own storage.
Gemini 3 Pro Image at a glance
The facts that decide whether Gemini 3 Pro Image fits the job.
- Model ID
- gemini-3-pro-image
- Output sizes
- 1K, 2K, 4K
- Aspect ratios
- 10 (9:16 to 21:9)
- References
- Up to 14 images
- Thinking
- Always on; optional thought summaries
- Billing
- Per image (Standard) or per token (Official)
Need 512px drafts, extreme aspect ratios or video input? See Gemini 3.1 Flash Image.
Before you move Gemini 3 Pro Image code over
The preview ID still works
Google retired gemini-3-pro-image-preview on June 25, 2026. Here it still runs, served by gemini-3-pro-image.
URLs, not inlineData
Send Gemini 3 Pro Image references as fileData URLs. Base64 inlineData is refused before anything is charged.
No Search grounding yet
Gemini 3 Pro Image requests with tools are refused for now.
One image per request
candidateCount is 1. Send several Gemini 3 Pro Image requests for a batch.
Why run Gemini 3 Pro Image on SeedRouter
Two ways to pay
Pick Standard for one flat price per Gemini 3 Pro Image image, or Official to pay for the tokens each render reports. Same key, same request body.
Failed requests are free
If a task ends failed, nothing is charged. You only pay for images that are delivered.
Credits never expire
Top up once and use the balance whenever you need it. No subscription.
Hosted results
Finished images come back as URLs on our CDN. Keep a copy of anything you need long term.
One key for every model
The same key and balance work for the other image models in the catalogue.
Other image models on SeedRouter
Same key, same balance, same task API.
Create and edit images with GPT Image 2 on SeedRouter. Use text prompts, reference images and transparent backgrounds, with a live playground and usage-based pricing.
View pricingGenerate and edit images with GPT Image 2.5 Flare or Sunburst on SeedRouter. Six quality steps up to max, transparent backgrounds, a live playground and pay-as-you-go pricing.
View pricingUse Gemini 3.1 Flash Image (gemini-3.1-flash-image) with Google's generateContent body: 512px to 4K, 14 aspect ratios, 14 references. Pay per image.
View pricingNano Banana 2.1, Google's update to Nano Banana 2: create and edit images with sharper text, 1K to 4K, 14 aspect ratios and 14 references, one price per image.
View pricingUse Gemini 3.1 Flash Lite Image (gemini-3.1-flash-lite-image), Google's fastest and cheapest image model, with its generateContent body. 1K output.
View pricingGemini 3 Pro Image guides
Practical guides to choosing, pricing and integrating the model.
Compare Nano Banana Pro API prices for 2K images, check token and per-image billing, and estimate a batch budget using current SeedRouter pricing.
WaveSpeed AI API pricing per model for Seedance, Nano Banana, GPT-6 and Claude, its $1 free credit and account levels, and where the same models cost less.
Atlas Cloud API pricing for Seedance, Nano Banana Pro and GPT Image 2 from its own calculator, what its free credits cover, and a cheaper option.
fal API pricing per model for Seedance 2.5, Seedance 2.0 and Nano Banana, read from fal model pages, with a tier-by-tier comparison against SeedRouter prices.
Higgsfield API pricing per model, how its launch discounts and $15 credit work, whether the API is free, and where Seedance and Nano Banana Pro cost less.
Kie API pricing per model for Seedance, Nano Banana and GPT Image, converted from Kie credits to dollars, next to a Kie API alternative with lower rates.
A Krea AI alternative for API users: Krea API prices for Seedance 2.0, Fast, Mini and Nano Banana next to a cheaper option, plus Krea plans and free units.
Pollo AI API pricing per model, the $80 minimum top-up, whether the Pollo API is free, and where Nano Banana Pro and Nano Banana 2 cost less per image.
Looking for a Replicate alternative? Compare Replicate prices for Seedance 2.0 and Nano Banana with SeedRouter, what is free on each, and how billing differs.
GPT Image 2.5 vs Nano Banana Pro: how the two image models compare on text, editing, reference images, output sizes and request shape, and which to pick.
Give Claude Code a Nano Banana Pro skill: a small script that submits the request, polls the task and saves the image, plus a SKILL.md you invoke by name.
How to write Nano Banana Pro prompts: a five-part structure, exact text in images, roles for reference images, edits that keep the rest, and 2K or 4K output.
Is Nano Banana free? What the Gemini app and the Gemini API include for Nano Banana 2 and Nano Banana Pro without paying, and when you start paying.
Use the Nano Banana 2 API step by step: create a key, send a generateContent request, poll the task, add reference images, and let a coding agent run it.
Compare Nano Banana 2, Nano Banana Pro and Nano Banana 2 Lite on image size, aspect ratios, references, thinking and billing, and pick one per job.
Current Nano Banana API pricing for Nano Banana 2, Pro and 2 Lite: flat per-image and token billing, cost by resolution, and when each billing fits.
Gemini 3 Pro Image FAQ
Is Gemini 3 Pro Image the same as Nano Banana Pro?+
Yes. Nano Banana Pro is Google's product name for Gemini 3 Pro Image, and the model ID is gemini-3-pro-image. The Nano Banana Pro page covers the same model from a product angle.
Is gemini-3-pro-image-preview still available?+
Google shut the preview ID down on June 25, 2026. On SeedRouter gemini-3-pro-image-preview is still accepted and runs on gemini-3-pro-image, so old code keeps working while you switch the ID.
How much does Gemini 3 Pro Image cost?+
An image costs $0.01875 on Standard at any size. On Official you pay for the tokens each render reports; a typical 1K or 2K render comes to about $0.10304. The pricing section shows every size at the live rate.
Is Gemini 3 Pro Image free?+
Google's pricing page lists no free tier for Gemini 3 Pro Image on the Gemini API. On SeedRouter you top up once, credits never expire, and failed requests are not charged.
Do I need to change my Gemini request?+
Keep contents, systemInstruction, safetySettings and generationConfig as they are. Add a model field, send it to /v1/images/generations, and read the Gemini 3 Pro Image result from the task.
How do I get an API key for Gemini 3 Pro Image?+
Sign in, top up any amount and create a key on the API keys page. The same key works for Gemini 3 Pro Image and every other model here.
Can I turn thinking off on Gemini 3 Pro Image?+
No. Gemini 3 Pro Image always thinks before it draws. You can ask for a thought summary with includeThoughts, and thinking tokens are billed on Official only.
Gemini 3 Pro Image or Gemini 3.1 Flash Image?+
Use Gemini 3 Pro Image for complex instructions and professional assets. Gemini 3.1 Flash Image is faster, adds 512px output and extreme aspect ratios, and reads video and PDF references.
Send your first Gemini 3 Pro Image request
Try Gemini 3 Pro Image in the Playground, then paste the same JSON into your code.