Recently, the best frontier AI model ever made available to the public came out: Fable 5.2. It beats virtually every benchmark, in some cases very distinctly above its competitors, and it’s the best so far at visual design. Some user put it to the test, along with other AI technologies, to write the lyrics of a song and eventually direct the Suno song as a music video. The result is one of the most impressive things that I’ve seen involving AI, so I want to record it here to return to it.
Sadly, I can’t find any YouTube video or any way that would allow me to embed it here, so the link to the page that hosts that music video will have to do:
The page indicates everything that was involved in producing this fantastic music video. Different AIs were in charge of creating the entire song virtually from zero, with the human just coordinating.
Song
4:00 · Suno v6-wild, take S4b of 14, picked by anabology by ear · a theatrical soul shuffle at 117.45 BPM with a music box on the hook · lyric by Claude Fable 5.2
Edit
49 shots on a line table read off the song’s stems · 13 set plates painted in SVG, in code · a silhouette film drawn and animated in code
GPT Image 2.6 Sunburst
61 stills, $7.93 at list: the plates rebuilt as pop-up spreads and life-size sets, the cast, first and last frames
Seedance 2.5
84 takes on Higgsfield’s API and Replicate over two versions, $202.23 at list · v1’s 23 more made in Higgsfield’s app
Made in
two passes: v1 in 5 h 08 min from one brief, v2 re-shot on the song anabology picked in 3 h 01 min · about $210 at list
