Nova Canvas 4,194,304 pixel cap and 16 px rule vs Sume image_size
Nova Canvas output sides must divide by 16 and total under 4,194,304 pixels. GPT Image 2.5 on Sume has its own custom-size rules. A side-by-side check.
Both Nova Canvas and GPT Image 2.5 on Sume refuse sizes that look normal, such as 1080 by 1350. They refuse them for different reasons, so a size you validated for one will not carry over.
| Rule | Nova Canvas | GPT Image 2.5 `image_size` on Sume |
|---|---|---|
| Side range | 320 to 4096 px (input and output) | Max edge 3840 |
| Aspect ratio | 1:4 to 4:1 | At most 3:1 |
| Pixel budget | Under 4,194,304 total | 655,360 to 8,294,400 total |
| Divisible by 16 | Output sides yes, input sides no | Yes |
Where the surprises are
On Nova Canvas, an input side does not need to be a multiple of 16 but an output side does. On Sume, 1080 by 1350 fails on GPT Image 2.5 for the same multiple-of-16 reason. See GPT Image 2.5 1080x1350 for the fix.
Check a size before you send it
This function tests a width and height against the Sume GPT Image 2.5 rules from the table. It is local, so a bad size never costs a request.
def gpt_size_ok(w: int, h: int) -> list[str]:
errs = []
if w % 16 or h % 16:
errs.append("sides must be multiples of 16")
if max(w, h) > 3840:
errs.append("max edge is 3840")
if max(w, h) / min(w, h) > 3:
errs.append("aspect ratio must be at most 3:1")
if not 655_360 <= w * h <= 8_294_400:
errs.append("pixels must be 655,360 to 8,294,400")
return errs
for size in [(1080, 1350), (1088, 1360), (1536, 864)]:
print(size, gpt_size_ok(*size) or "ok")Sources
Related posts
More in Comparisons
- Nova Canvas IMAGE_VARIATION similarityStrength vs Sume n and reference
Nova Canvas IMAGE_VARIATION takes 1 to 5 images and a similarityStrength of 0.2 to 1.0. Sume has no strength field; here is how to get variations.
- Nova Canvas inpainting mask rules vs a Sume reference edit
Nova Canvas inpainting wants a pure black and white mask, same size as the input. What to build, and how Sume's edit call differs.
- Nova Canvas maskPrompt: edit by naming the region, and the Sume way
Nova Canvas accepts a maskPrompt in place of a maskImage. Sume edits work from the prompt alone on most models. How to word the region and test the result.
- Nova Canvas outpainting DEFAULT vs PRECISE, and a Sume ratio edit
Nova Canvas outPaintingMode DEFAULT can leave a halo; PRECISE follows the mask. Here is how to extend a photo on Sume, and how to test the seam.
Written by Sume