WHAT NEW IN X10.31 SAT

AI collections from X · 22 collections · 62,909 posts

chatgpt-imagesIMAGE

Extracting one image from a grid with plain-language prompts in GPT Image 2.5

A test shows GPT Image 2.5 understands casual instructions like "second row from the top, second from the left" to pull a single image out of a grid.

ミサキ
@misakin883Archived 2026.09.14Originally posted 2026.09.14
Extracting one image from a grid with plain-language prompts in GPT Image 2.5 (image 1 of 4)
Extracting one image from a grid with plain-language prompts in GPT Image 2.5 (image 2 of 4)
Extracting one image from a grid with plain-language prompts in GPT Image 2.5 (image 3 of 4)
Extracting one image from a grid with plain-language prompts in GPT Image 2.5 (image 4 of 4)
ミサキ @misakin883Originally posted 2026.09.14
Open original on X

グリッドから1枚だけ抜き出すプロンプト、GPT Image 2.5は雑に言っても通じますね。 青○の2か所を、それぞれこの言い方で取り出せました。 「上から2番目で左から2番目の画像を1枚で抽出して」 「縦3*横2の画像を1枚で抽出して」 普段の喋り方のままで伝わるの、助かります😊

This post quotes another post: x.com/misakin883/status/2099337288673325348

x.com/misakin883/status/2099351178077413642Archived 2026.09.14 19:43