OracleZoom zooms 256x into photos by writing its own prompt at each 4x step

multimodalart · x · 2026-09-14

HuggingApps' OracleZoom zooms 256× into a photo in 4× steps, approaching super-resolution by having the model write its own prompt at every zoom level and re-render the image.

The twist inverts the usual workflow: instead of a human prompting the model, the model decides at each level what details to hallucinate, iteratively turning low-resolution input into a coherent enlarged image.

Original post →

More from Multimodal

Multimodal channel →