Nano Banana 2.1 at 8:1 and 4K: 12288x1536, 18.87 megapixels
Google's size table puts Nano Banana 2.1 8:1 at 4K at 12288x1536, about 18.87 MP, larger than a 4096x4096 square. Sume lists 8:1 and 4K at $0.20 flat.
The largest Nano Banana 2.1 output in Google's own size table is the 4K tier at 8:1 or 1:8: 12288 by 1536 pixels, or 1536 by 12288. That is about 18.87 megapixels, more than the 16.78 megapixels of a 4096 by 4096 square. Sume lists the 8:1 ratio, the 1:8 ratio and the 4K tier for Nano Banana 2.1, and bills the 4K tier at $0.20 per image whatever the ratio.
That makes the extreme ratios cheap per pixel. At a flat $0.20 the 8:1 image costs about $0.0106 per megapixel, while the square costs about $0.0119. The difference is small but it runs in the direction people do not expect.
Sizes from the Google table
The sizes below are copied from the 4K column of the Nano Banana 2.1 table on Google's image generation page (read 2026-10-09). Megapixels are width times height divided by one million. The last column divides Sume's $0.20 4K price by the megapixels.
| Ratio | 4K size | Megapixels | Sume $ per MP at $0.20 |
|---|---|---|---|
| 1:1 | 4096x4096 | 16.78 | $0.0119 |
| 16:9 | 5504x3072 | 16.91 | $0.0118 |
| 21:9 | 6336x2688 | 17.03 | $0.0117 |
| 4:1 | 8192x2048 | 16.78 | $0.0119 |
| 8:1 | 12288x1536 | 18.87 | $0.0106 |
| 1:8 | 1536x12288 | 18.87 | $0.0106 |
What Sume does and does not promise
Sume's model descriptors list 4K among the resolution values and 8:1, 1:8, 4:1, 1:4 and 21:9 among the ratios for Nano Banana 2.1. The Sume docs do not state the pixel size of a returned image, so the Google table is a guide to what the underlying model produces and not a Sume guarantee. Read the width and height of the file you get back before you build a layout on it.
Also keep the delivery in mind. 4K requests are among the slow configurations that can degrade to a 202 job envelope, in which case you poll GET /v1/jobs/:id/status and read the result from /result.
Compute the cost per megapixel yourself
The loop below reproduces the table from the sizes. It only does arithmetic, so it runs as is.
sizes = {
"1:1": (4096, 4096),
"16:9": (5504, 3072),
"21:9": (6336, 2688),
"4:1": (8192, 2048),
"8:1": (12288, 1536),
}
PRICE_4K = 0.20 # Sume Nano Banana 2.1, 4K, per image
for ratio, (w, h) in sizes.items():
mp = w * h / 1_000_000
print(f"{ratio}: {w}x{h} = {mp:.2f} MP, ${PRICE_4K / mp:.4f} per MP")When to use it
A 12288 pixel strip is useful for a long panoramic header, a scrolling timeline graphic or a printed banner. For web use, resize it to the width you serve, since a browser will not need 12,000 pixels. If the strip is for a screen of 3840 by 480, an 8:1 image at 2K already has more pixels than the screen shows. See the 8:1 and 21:9 comparison for the other models.
One more planning point: the 8:1 and 1:8 shapes are the only ones where the long edge passes 10,000 pixels, so image viewers and some design tools may refuse the file or open it slowly. Test the full-size file in the tool your team uses before you commit to the tier. If it opens poorly, generate at 2K, which Sume bills at $0.15, and compare the two crops side by side at the size you will actually publish.
Sources
Related posts
More in Models
- Nano Banana 2.1 9:16: 768x1376, 1536x2752 or 3072x5504 for stories
The three 9:16 sizes Nano Banana 2.1 outputs and how to reach 1080x1920 for Stories and Reels covers with a 15 px vertical trim.
- nano-banana-2 still works but runs as 2.1: check the stored job id
Sume retired Nano Banana 2: google/nano-banana-2 runs as Nano Banana 2.1 and the job stores the 2.1 id. A Python check and what you pay.
- Nano Banana 2 to 2.1 on Google: 4K falls from $0.151 to $0.113
Google's price per image falls about 50% at 1K and 2K and 25% at 4K from Nano Banana 2 to 2.1. Sume runs nano-banana-2 requests as 2.1 at its own 2.1 rates.
- Nano Banana Pro 4:5 at 1K is 928x1152, not 1080x1350: the 1.16x gap
Sume's docs say Nano Banana Pro sends 4:5 at about 928x1152 at 1K, and exact 1080x1350 comes from a post-step. That is 1.16x wide and costs $0.1875 per image.
Written by Sume