Nano Banana 2 Lite API
AI ImageThe Nano Banana 2 Lite API generates and edits 1K images on OmniaKey for $0.019 per call.
- Per successful call$0.019
- OperationsText to image
- Image edit
- View tool pricing and billing rules
Your result will appear here once you run the model.
01
Where it fits
The Nano Banana 2 Lite API generates and edits 1K images on OmniaKey for $0.019 per call.
Try Lite for drafts and simple edits. Nano Banana 2 costs the same at 1K and also offers 2K and 4K.
02
What it does
Text to image
gemini-3.1-flash-lite-image
Turn a text prompt into a new image.
Image edit
gemini-3.1-flash-lite-image
Change an existing image from a prompt and reference input.
03
In depth
Nano Banana 2 Lite is Google's small Gemini image model on OmniaKey under the id gemini-3.1-flash-lite-image, marked for text-to-image and image-edit. It is the only id in this family whose catalog entry carries no size matrix: one rate, one output size of 1K, and Google's model page excludes 2K and 4K outright. At 1K it sits on the same listed tier as Nano Banana 2, so the choice between them is about size range and editing depth rather than a discount.
Google positions Lite as the lightest member of the family, and says it is not optimized for several reference images or for sequential multi-turn edits. Take that at face value: a colour swap, a background change or a thumbnail batch fits, while a chain of brand-sensitive revisions does not. The id is also easy to mistype, because gemini-3.1-flash-lite without the -image suffix is a text model rather than this one.
Platform behaviour is worth a line of its own. A 1K label names an output class rather than a square, so send aspect_ratio and image_size together and trust the delivered dimensions; the body cap of 1 MiB applies to edits, so a camera file needs downsizing first; and reusing an Idempotency-Key with a changed body returns 409, which a retry loop has to handle.
Thumbnails and grid drafts
Feed category pages and list views from Lite: one fixed size, one flat rate, and enough detail for a card that is read at a few hundred pixels. Generate the whole grid, then re-run only the frames a customer actually opens on a model with a larger canvas.
Single-step edits
A colour swap or a backdrop change on one product frame sits inside its remit, and Google's caution about sequential edits marks the edge of it. Make the change in one task, keep the source under the 1 MiB body cap, and compare the label and the silhouette against the original.
Settling a prompt cheaply
Use Lite to fix the wording before a larger run: four variants at 1K are enough to see which phrasing survives, and that phrasing then goes to Nano Banana 2.1 at 2K or 4K. Composition decisions carry over even though the pixels do not, so keep the prompt text next to the final file.
04
Limits to plan around
- No 2K or 4K: Google's page for this id excludes both, so a large deliverable has to be generated elsewhere rather than enlarged from a 1K frame.
- Google says the model is not optimized for several reference images or sequential multi-turn edits, so a long chain of brand revisions belongs on a full model.
- The id sits one word from a text model: gemini-3.1-flash-lite without the -image suffix answers as text, so a task body carrying the wrong id fails in a confusing way.
- One rate, one size and one image per call: a bigger canvas is not on offer here, which makes volume the only lever on a large job.
05
Compare nearby models
| Model | Best for | Operations | Listed price |
|---|---|---|---|
| Nano Banana 2 LiteThis page gemini-3.1-flash-lite-image | Thumbnails, prompt drafts and single-step edits where a 1K frame is the whole deliverable. | Text to image · Image edit | $0.019 per call |
| Nano Banana 2 gemini-3.1-flash-image | The same 1K tier with 2K and 4K above it, plus a 512 option this model does not list. | Text to image · Image edit | $0.019–$0.039 per call |
| Nano Banana 2.1 gemini-nano-banana-2.1 | Reference-heavy edits and 4K output, with a thinking level that starts higher by default. | Text to image · Image edit | $0.019–$0.039 per call |
06
Sources and verification
Guide updated 2026-10-27.
07
Call it
Media models run asynchronously. Create a task, poll it until it finishes, then download the signed asset URL.
# 1. Create the job. 202 Accepted; the body carries the task id,
# the pre-charged price, and a poll_url. Reuse one Idempotency-Key
# across retries so a timeout never starts a second paid job.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks \
-H "Authorization: Bearer $OMNIAKEY_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: gemini-3.1-flash-lite-image-001" \
-d '{
"model": "gemini-3.1-flash-lite-image",
"type": "image",
"operation": "generate",
"input": {"prompt": "a paper boat on a quiet lake at sunset"},
"parameters": {"aspect_ratio": "1:1"}
}'
# 2. Poll GET /v1/media/tasks/{id} until status is
# succeeded | failed | canceled. Progress carries percent and stage.
curl --fail-with-body https://api.omniakey.com/v1/media/tasks/TASK_ID \
-H "Authorization: Bearer $OMNIAKEY_API_KEY"
# 3. Download assets[0].url before expires_at, or poll again for a
# fresh signed URL. Failed and canceled jobs are refunded.
curl -sS -o out.png "$(jq -r '.assets[0].url' result.json)"08
Pricing
Image jobs are billed per successful call, not per token. Failed or canceled jobs are refunded automatically.
$0.019per callSave 43%
More Google image models
Related articles
Nano Banana 2 Lite vs Nano Banana 2
Choose by resolution, editing needs and cost per usable image.
Read articleSeedream 5.0 Pro vs Nano Banana Pro: Same-Prompt Test and Price
Five shared prompts, two models, measured times and the price per 2K image.
Read articleGPT Image 2.5 vs Nano Banana 2.1: Same-Prompt Test and Price
Four prompts, three image models, one day: what each made and what it really cost.
Read articleQwen Image 3 vs Nano Banana 2.1: Same-Prompt Test and Price
Four prompts, two image models, one day: what each made and what it cost.
Read articleFrequently asked questions
How do I call Nano Banana 2 Lite on OmniaKey?
POST to /v1/media/tasks with the model id, type "image", and operation "generate" or "edit". Reuse one Idempotency-Key across retries so a timeout never starts a second paid job.
What does Nano Banana 2 Lite cost?
Nano Banana 2 Lite bills $0.019 per call. Creation pre-charges that amount from your prepaid balance, and failed or canceled jobs are refunded automatically.
Why does the API return a task id instead of the image?
Provider latency runs from seconds to about a minute, so OmniaKey normalizes every image route into one asynchronous lifecycle. Poll GET /v1/media/tasks/{id} until the status is succeeded, then read the signed asset URL.
Is Lite cheaper than Nano Banana 2?
Lite is billed per call at the rate in the price table on the model page, and so is Nano Banana 2; at 1K the two share one amount, so what separates them is range. Lite offers a single size while Nano Banana 2 adds 2K, 4K and a 512 option. A failed task returns its reserved credit, so a lost thumbnail costs nothing.
What kind of edit is Lite good for?
Simple local changes: a colour swap, a background replacement, one object removed. Google's guide says the model is not optimized for several references or for a chain of sequential edits, so keep each task to a single change and treat a fourth round as a signal to move up.
Which parameters does a request take?
aspect_ratio and image_size, the Gemini pair, rather than the size and quality fields that GPT Image uses. Send image_size as 1K, keep the whole body under 1 MiB when sending an edit, and mint a fresh Idempotency-Key whenever the body changes.
Why did my task return 409?
Because the Idempotency-Key was reused with a different body. Reuse a key only to repeat an identical request; after editing the prompt, the aspect ratio or the size, mint a new one, and read a 409 as a signal to fix the key rather than to retry blindly.
Can I upscale a Lite frame to 4K?
Only as a separate processing step, and the result will not carry detail the model never drew. If the deliverable is 2K or 4K, generate it on Nano Banana 2 or 2.1; upscaling a 1K file is a fallback for a preview, not a substitute for a native render.