SciFig AI

GPT Image 2 for research visuals

Use GPT Image 2 for structured, label-aware scientific drafts

GPT Image 2 is the broadest starting point in SciFig when you need prompt precision, readable text, flexible aspect ratios, and the option to revise an existing image.

Scientific illustration showing an antibody and antigen binding interface
GPT Image 2Precise prompts, clean text
Structured scientific diagram of the glycolysis pathway

Ordered biochemical pathway

Use explicit step order, branch logic, and legend instructions to test layout control.

Where this model fits well

  • Graphical abstracts with a clear left-to-right research story
  • Mechanism diagrams that need concise, readable labels
  • Iterative image-to-image refinement from a sketch or prior draft
  • Layouts that need explicit aspect-ratio and resolution choices

What to plan around

  • Readable text is not the same as correct text; every label still needs review.
  • Complex molecular structures, pathways, scales, and causal arrows can be invented or misplaced.
  • Some aspect-ratio and scene combinations restrict the available resolution in the generator.
  • The first output is best treated as a composition draft, not a submission-ready figure.
Model

Current Kie integration limits

These values describe the limits exposed by SciFig’s current Kie-backed model registry, not every capability of the underlying model.

GPT Image 2

Precise prompts, clean text

Generation modes
Text to image / Image to image
Resolution
1K / 2K / 4K
Kie/SciFig prompt limit
20,000 characters
Kie/SciFig reference-image limit
16 images
Auto ratio
Supported
Aspect ratios
auto, 1:1, 3:2, 2:3, 4:3, 3:4, 5:4, 4:5, 16:9, 9:16, 2:1, 1:2, 3:1, 1:3, 21:9, 9:21

Provenance: SciFig reads these limits from its Kie integration and shared model registry. The public panel may apply lower product-level input limits, and compatible resolutions can vary by scene and aspect ratio.

Test prompt

Scientific structures worth testing

These existing SciFig examples show useful information structures, not verified GPT Image 2 benchmarks. Use the prompt to test the current model and compare your own result.

Scientific illustration showing an antibody and antigen binding interface

Labeled molecular interaction

Test whether the model can organize a main structure, detail inset, and concise legend.

Scientific figure of antibody-antigen binding with a full antibody overview, enlarged binding-site inset, concise labels for variable and constant regions, and a small interaction legend; clean white background, publication figure layout

Structured scientific diagram of the glycolysis pathway

Ordered biochemical pathway

Use explicit step order, branch logic, and legend instructions to test layout control.

Structured overview of glycolysis from glucose to pyruvate, numbered vertical steps, ATP and NADH inputs and outputs, compact legend, clean scientific infographic on white background; all labels must be checked by the author

Scientific diagram of a spiral galaxy with labeled orbital paths

Diagram with labels and insets

A useful test for spatial hierarchy, annotations, and multiple coordinated panels.

Scientific diagram of a spiral galaxy with major arms labeled, three orbital paths overlaid, a side-view inset, scale bar, and compact legend; restrained blue palette, white background, clean educational figure

Generate with GPT Image 2 selected

Start with a sample or describe the system, visual hierarchy, labels, and reading direction in your own words.

0/2000

Ilmainen esikatselu, kuva luotu osoitteessa 1K.Kirjaudu sisäänSen jälkeen voit avata sivustot 2K / 4K ja tallentaa selaushistorian.

Compare models and continue creating

Nano Banana family

Compare fast generation and multi-reference editing options.

Z-Image

Explore a focused photorealistic text-to-image option.

Graphical Abstract Generator

Use a workflow built around research-story structure.