About This Prompt

A detailed prompt designed to test the Japanese text rendering and layout capabilities of GPTs Image 1.5 versus Nano Banana Pro. It specifies a 3-panel horizontal manga format, character consistency, and requires precise, uncorrupted Japanese text within speech bubbles, comparing the strengths of each model.

Prompt Details

ID: 2645

Requires Reference Images: No

Sample Images

Full Prompt

[Format] Horizontal 3-panel manga (left to right). Anime style + pop. Bright color scheme. Clean lines. Simple office background.
[Characters] Emma (20s female, black bob, white shirt + navy jacket, expressive face). Consistent appearance across all panels.
[Text] Japanese text displayed clearly and legibly. No typos, garbled characters, or broken text. Use the following text in the speech bubbles "exactly as written":

Panel 1 (Left):
Emma is explaining while pointing.
Speech bubble text (exact match):
"{argument name="Panel 1 Text" default="ChatGPT Images is good at illustrations, anime, and 3D expressions"}"

Panel 2 (Center):
Emma is doing design work (layout and typesetting).
Speech bubble text (exact match):
"{argument name="Panel 2 Text" default="Nano Banana is good at Japanese text and design"}"

Panel 3 (Right):
Comparison panel. Split the screen left/right, with headings "ChatGPT Images" on the left and "Nano Banana" on the right.
Include checklists for each (in Japanese, short, and readable).
Left (ChatGPT Images): 'Illustration', 'Anime', '3D Expression'
Right (Nano Banana): 'Japanese Text', 'Design', 'Layout'
Also, place a large "VS" in the center to compare the two compositions.

[NG] No unnecessary text, arbitrary English additions, watermarks, signatures, unreadable text, or broken Japanese.
[Image Quality] High resolution, sharp, suitable for SNS posting.

---