Art, video & music

Pictures drawn by two LLMs emitting raw RGB values as text

Curious whether a language model could draw like a pixel-predicting image model, Logan Bolton, an MS student at NYU Courant, gave Claude Opus and Fable the same prompts and had them output every pixel's RGB values into a .txt file. Opus came out looking much better, to their surprise.

Made with

Claude Opus, Claude Fable

Role

MS student, NYU Courant

Posted

17 September 2026

Logan Bolton@septisum

I was curious about what it would look like for an LLM to do something similar, so I gave Opus and Fable the same prompts and had it generate the image by just outputting raw RGB values in a .txt file

Opus looks... way better? I was not expecting this

17 Sep 2026View on X