Nano Banana API
AI ImageNano Banana is Google's original Gemini 2.5 Flash Image route on OmniaKey, supporting text-to-image generation and reference-based edits through the media API.
- 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
Nano Banana is Google's original Gemini 2.5 Flash Image route on OmniaKey, supporting text-to-image generation and reference-based edits through the media API.
Use Nano Banana for existing Gemini 2.5 image workflows. Nano Banana 2 has the same 1K price and offers larger outputs.
02
What it does
Text to image
gemini-2.5-flash-image
Turn a text prompt into a new image.
Image edit
gemini-2.5-flash-image
Change an existing image from a prompt and reference input.
03
In depth
Nano Banana is Google's original Gemini 2.5 Flash Image model on OmniaKey, listed as gemini-2.5-flash-image with text-to-image and image-edit. Nano Banana is Google's product name for Gemini's native image generation and editing, and this is the oldest member of that family here: the catalog gives it a single per-call rate with no size tiers and no preview alias beside it.
Its most useful job in our own work had nothing to do with winning a comparison. We used it to make the source pictures that other models then edited, so no model judged its own output, and a neutral frame is exactly what a legacy model is good at. It is also the id that existing Gemini 2.5 code already calls.
Two limits shape the decision. The catalog lists no 2K or 4K step for this id, so a large deliverable means a different model; and the improvements Google shipped later, on text rendering, references and multi-turn edits, sit on the newer ids. Keep this one for continuity and for source material, and start new briefs on Nano Banana 2.1 or Pro.
Neutral frames for edit tests
When several models have to edit the same picture, make that picture here. Our October six-test round did exactly that, so no model was asked to improve its own output and the comparison stayed honest. Keep the source plain, settle the light and the composition once, and reuse the file everywhere.
Existing Gemini 2.5 code
Clients written against the 2.5 generation keep working without a rewrite: the same task shape, the same edit operation, one rate per call. Pin the id in configuration rather than hard-coding it, so a later move to gemini-nano-banana-2.1 becomes a settings change plus a re-run of your sample set.
Plain product and lifestyle frames
Simple briefs still come out usable: one object, daylight, a clean background, no dense type. Skip it for posters, price columns or a group of people who must stay recognisable, and spend a newer model on those tasks instead of stretching this one.
04
Limits to plan around
- The catalog lists a single per-call rate and no size tiers for this id, so read the delivered dimensions rather than assuming a 2K or 4K canvas.
- It is the oldest member of the family here: the text, reference and multi-turn work Google shipped later belongs to the newer ids, not to this one.
- Google says generated images carry a SynthID watermark, which marks provenance and says nothing about whether the picture is correct.
- An edit takes the source inline inside a 1 MiB body, so a full-size camera file needs resizing before the task can be created.
05
Compare nearby models
| Model | Best for | Operations | Listed price |
|---|---|---|---|
| Nano BananaThis page gemini-2.5-flash-image | Existing 2.5 workflows and neutral source frames, at one listed rate per call. | Text to image · Image edit | $0.019 per call |
| Nano Banana 2.1 gemini-nano-banana-2.1 | New work with exact lettering, ten object references and output from 1K to 4K. | Text to image · Image edit | $0.019–$0.039 per call |
| Nano Banana Pro gemini-3-pro-image | Complex layouts and portrait edits where the top listed tier buys physical detail. | Text to image · Image edit | $0.039–$0.069 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-2.5-flash-image-001" \
-d '{
"model": "gemini-2.5-flash-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 51%
More Google image models
Related articles
Gemini image API relay guide
Choose the right API contract, generate an image, and recover safely when a request times out.
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 articlenano banana 2.1 vs pro
Six jobs, the same prompts, three models: what each made and which to pick.
Read articleFrequently asked questions
How do I call Nano Banana 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 cost?
Nano Banana 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 Nano Banana the same as Nano Banana 2?
No. Nano Banana is Google's product name for Gemini's native image generation, and this id, gemini-2.5-flash-image, is the oldest family member on the platform. Nano Banana 2 uses gemini-3.1-flash-image and Nano Banana 2.1 uses gemini-nano-banana-2.1, each a separate model with its own behaviour.
Can this model produce a 4K image?
The catalog lists no size steps for this id, so plan around the size it returns instead of a canvas you pick. If a 4K file is the goal, Nano Banana 2 and 2.1 both list 1K through 4K, and Pro adds a higher top tier of its own above them.
How is billing counted on this model?
One image per call at a single rate, with no size steps to choose between, so repeating the same brief costs the same each time. The current rate is in the price table on the model page. A failed task returns its reserved credit, and an expired asset link can be refreshed by polling the same task again.
Why keep an older image model in the stack?
For continuity and for source material: prompts that already pass review keep working, and a neutral frame for an edit comparison does not need the strongest model. New briefs with dense type, several references or a large canvas belong on a newer id.
Does it add a watermark to the picture?
Google says generated images include a SynthID watermark, and this model belongs to that family. The mark is provenance metadata rather than a visible logo, and it says nothing about whether a label, a count or a face came out right.