String Resolution GPT Image 2.0 - GPT Image 2.0 Resolution Selector
Node Purpose: String Resolution GPT Image 2.0 outputs the nearest GPT Image 2.0 resolution tier label (1k / 2k / 4k) for each input image based on image area. If no image is connected, it passes through the manually selected label. It is designed to pair with GPT Image 2.0.
Inputs
| Name |
Port |
Type |
Default |
Range |
Description |
selection |
- |
COMBO |
1k |
1k / 2k / 4k |
Manual GPT Image 2.0 resolution label used when image is not connected. |
image |
optional |
IMAGE |
- |
- |
Input image batch used to infer the nearest resolution tier per frame. |
Outputs
| Name |
Type |
Description |
string |
STRING |
Resolution label string; batched inputs are joined by newline in batch order. |
Tiers and target areas
| Tier |
1:1 official size |
Target area (pixels) |
1k |
1024x1024 |
1,048,576 |
2k |
2048x2048 |
4,194,304 |
4k |
3840x2160 (1:1 is 2880x2880) |
8,294,400 |
Features
- GPT Image 2.0 tiers:
1k / 2k / 4k, using each tier's official 1:1 size area as the target.
- Area-based matching: compares image area against the tier target areas and picks the nearest in log space.
- Manual fallback: returns the selected label directly when
image is empty.
- Batch support: outputs one resolution label per image in the batch.
Typical Usage
- Convert image size into a resolution label before GPT Image 2.0 nodes that expect a resolution string, and connect it to the API node's
resolution input.
- Keep manual control by leaving
image unconnected and selecting a fixed label.
Notes & Tips
- Matching is based on total pixel area, not exact width/height ratio.
- The target areas match
GPTImage20.MAX_PIXELS_BY_RESOLUTION in 1hewOffice (1k=1024², 2k=2048², 4k=3840x2160, the same area as the 4k 1:1 size 2880x2880) item by item; that source belongs to 1hewOffice and its code is not in this repository.
- This node only outputs the string label and does not resize the image.
