The 4K checkbox costs one credit.
1K is 3 credits, 2K is 4, 4K is 5. So the jump from a screen-sized draft to a print-sized master is a third more, not three times more — and this is the one place in my image stack where I stopped thinking about price and started thinking about what the deliverable actually is.
I noticed it on Callirra, where the ladder is three, four and five credits and the checkbox is a single control.
The ladder
| Resolution | Credits | Dollars | What it is for |
|---|---|---|---|
| 1K | 3 | $0.03 | drafting, social, thumbnails |
| 2K | 4 | $0.04 | web hero, most editorial |
| 4K | 5 | $0.05 | print, crop headroom, large formats |
For reference, Google's published image-output equivalents for the same model are about $0.0336 at 1K, $0.0504 at 2K and $0.113 at 4K. The 4K line is the interesting one: the step from 2K to 4K is where the reference rate more than doubles, and where a reseller's ladder either helps you or does not.
The batch math that changed my mind
I used to draft at 1K and upscale whichever frame survived. Then I ran the numbers on a 100-image deliverable:
| Plan | Credits | Dollars |
|---|---|---|
| 100 × 4K | 500 | $5.00 |
| 100 × 4K, but 30 drafts at 1K first, 10 finishes | 90 + 50 = 140 | $1.40 |
| 100 × 1K, upscaled later | 300 | $3.00 |
The middle row is the workflow: thirty cheap explorations at 1K (3 each), then ten 4K finishes (5 each). It is cheaper than rendering a hundred frames at 1K and upscaling — because you are not paying to render the frames you throw away at high resolution, and you are not inventing pixels afterwards either.
That is the whole argument for a ladder: the cheap tier is for deciding, the expensive tier is for delivering, and the expensive tier is only one credit more.
The panorama exemption
Two things only exist above 1K on this model: fine detail, and the extreme aspect ratios. Ratios like 1:4, 4:1, 1:8 and 8:1 are available at 2K and 4K. If your deliverable is a banner or a vertical strip, the resolution choice is made for you.
For everything else — 1:1, 3:2, 2:3, 16:9, 9:16 — 1K is a legitimate preview of the composition, and 4K changes only how much of the frame survives a crop.
What the request looks like
curl https://api.callirra.com/v1/images \
-H "Authorization: Bearer $CALLIRRA_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "nano-banana-2-1",
"prompt": "a premium editorial launch poster with exact readable headline…",
"resolution": "4K",
"aspect_ratio": "2:3",
"output_format": "png",
"references": ["https://cdn.example/ref-01.png"]
}'Two practical notes from the model's surface: up to 14 references, prompts up to 20,000 characters, and the JPG/PNG switch is an output format, not a transparency switch.
Where this goes wrong
Upscaling a 1K frame to 4K and calling it a 4K deliverable. You are inventing detail. It is fine for a background; it is not fine for a product label.
Paying for 4K on a frame that will be displayed at 600 px. The ladder is cheap, not free — 100 frames at 4K is still ten times the cost of 100 frames at 1K in the same month.
Assuming PNG gives transparency. On this model it does not; there are no background controls here.
Skipping the review at 100%. Text is stronger on this model than on most, and small type still slips. Review the 4K file, not the 1K preview.
Forgetting the resolution is per image, not per account. The ladder applies to every call; batching at 4K multiplies.
When 1K is the right answer
Social images, thumbnails, layout comps, anything that lives inside a 1080 px column, and any frame you have not decided about yet. On those, the extra credit buys nothing and the extra seconds buy less.
FAQ
Is 4K slower? Yes, modestly — larger output, same model.
Can I go from 2K to 4K by re-rendering? Yes, and it is a new charge (5 credits). Re-rendering at a higher tier is cleaner than upscaling.
Are the panoramic ratios 4K only? They are 2K and 4K.
Does the price change with aspect ratio? No. Resolution sets the price.
The current ladder and the Google comparison are on Callirra's Nano Banana 2.1 page. If you have a deliverable where the 1K-to-4K step did not pay for itself, that is worth writing down — the ladder is cheap enough that I stopped measuring, and that is exactly when a habit becomes a cost.