Gallery, prompt v1
Espresso Rag, with the left hand crossing over
One frozen prompt and every model that has run it, newest first.
The prompt, v1, frozen
Write, in LilyPond, 16 bars of café piano music called "Espresso Rag" in which the left hand crosses over the right hand for bars 9 to 12.
Every model, its first three runs.
One square a check, in this order:
- Piano staff
- Sixteen bars
- Left hand crosses
- Right hand keeps going
- Crossing marked
- Stems apart
passed, failed, undecided. A round mark is a check I judged by eye, and a diamond one a check a model judged.
| Claude Haiku 5.5 | |||
| Pareto 26.10 Preview | |||
| GPT-6 Luna | |||
| GPT-6.1 Sol |
What a run cost, against checks passed.
One dot a model, over its three runs above. Further right costs more. The gap from $0.01 to $0.1 is as wide as the gap from $0.1 to $1, so the cheap models are not squeezed into a corner.
How it is checked
LilyPond's own record of every note: its hand, staff, bar, pitch and stem, read while the file is engraved. A voice belongs to the hand whose staff it starts on, unless its notes are labelled with a hand's name. The script was written after hearing Sol's and Pareto's runs, so Luna was its first fair test. How good the music sounds is rated by ear.
When this saturates: Hand it an existing rag and ask for the same music with the hands swapped in bars 9 to 12, so every note has to stay where it is.
Failures seen here: incomplete (fails when the model leaves out something the brief asked for); unreadable (fails when the content may be right, but a reader can't make it out); false (fails when the model gets a fact wrong, such as a part in the wrong place or at the wrong size). The same kinds across every task.
The rules behind all of thisThis task's prompt, checks and any checker, on GitHub










