Mona Lisa SVG Challenge Opus 5.5 vs Sol 6.1
retained text
First prompt was very simple: *Make the mona lisa as an SVG image as best you can.* After realizing Sol clearly used a tracing technique based on the actual image, I tightened up the prompt: *Create the Mona Lisa as a single SVG file, drawn entirely from your own knowledge of the painting.* *Rules:* *- Do not search for, download, view, or use any reference image of the painting.* *- No tracing, vectorizing, or auto-converting any image, and no embedded raster data (<image>, base64, etc.).* *- Every shape must be authored by you, either directly or through code you write that does not take an image as input.* *- Keep the final file under 100 KB.* *- Canvas: 600 x 900 (viewBox="0 0 600 900").* *Aim for the most faithful and artistically convincing version you can: composition, proportions, the face and smile, the hands, the landscape, and the overall color and mood of the original. Gradients, filters, and blur are all allowed.* Both tools took quite a long time to generate this (over 15 minutes). Used High thinking level for both. I can't tell which one is derpier.
document fetched 2026-10-01 00:34 · canonical https://reddit.com/r/singularity/comments/1wtu7w8/mona_lisa_svg_challenge_opus_55_vs_sol_61