WHAT NEW IN X11.02 MON

AI collections from X · 22 collections · 65,263 posts

gpt-imageIMAGE

Generating SENSEX video images with the ChatGPT Go plan

The author shares an American comic-book style prompt template for batch-generating images for finance videos on the low-cost ChatGPT Go plan, and says the results look better than Grok's.

Dr. Mrunal Patel (Ph.D.)
@Mrunal_PatelArchived 2026.09.30Originally posted 2026.09.29
Generating SENSEX video images with the ChatGPT Go plan (image 1 of 4)
Generating SENSEX video images with the ChatGPT Go plan (image 2 of 4)
Generating SENSEX video images with the ChatGPT Go plan (image 3 of 4)
Generating SENSEX video images with the ChatGPT Go plan (image 4 of 4)
Dr. Mrunal Patel (Ph.D.) @Mrunal_PatelOriginally posted 2026.09.29
Open original on X

🌠 The images in my latest SENSEX video were generated using ChatGPT GO plan. I am mighty pleased with outputs for such low price! I have also used grok, but I find these chatGPT images more pleasing to the eyes. 👇🏻 Here is the prompt Create images using this consistent visual language: **STYLE** American comic-book editorial illustration; bold clean black ink outlines; halftone dot shading; dramatic but clear lighting; bright, peppy, saturated colors; pastel minimalistic backgrounds; strong visual contrast; polished professional comic/infographic aesthetic. **CHARACTERS** Keep people and symbolic characters large, expressive, clearly identifiable, and visually readable on a small mobile screen. Use bright, vivid clothing and skin tones. Exaggerated facial expressions and body language where appropriate. **COMPOSITION** Prioritize the exact layout and positioning specified in my prompt. Keep important characters, graphs, props, numbers, and symbols unobstructed. Use clear visual hierarchy and strong foreground/background separation. **INFOGRAPHIC ELEMENTS** When graphs, percentages, signs, charts, maps, or other visual data are requested, make them large, thick, highly legible, and visually prominent. Reproduce requested numbers and labels accurately in bold-thick fonts. Do not invent additional data or text. **PANELS** If I specify frames/panels, preserve the requested number, order, orientation, and relative proportions. Use clean panel separation unless I explicitly request borderless sections. **DEFAULT** No unnecessary text, labels, logos, captions, or decorative elements unless explicitly requested. Follow my scene description literally while using your creative judgment for minor visual details. #ToT (Training of the Trainers)

This post quotes another post: x.com/mrunal_patel/status/2104882251800645819

x.com/mrunal_patel/status/2104964228541190210Archived 2026.09.30 10:09