One Prompt, Ten Results
One prompt, ten results: the bedroom pop stress test
We ran June Okafor's bedroom pop recipe ten times without changing a word, rated every take in the open, and learned where the dice actually land.
SunoFetch Lab · 2026-08-12 · 8 min
The most common question in the lab inbox: if I run the same prompt twice, what changes? The honest answer is a distribution, not an anecdote. So we ran June Okafor's bedroom pop recipe ten consecutive times on the same model build, same day, and rated each take on our five dimensions before comparing notes.
[Bedroom pop, 96 BPM, G major] Soft drum machine, warbly cassette keys, one clean guitar line, close female vocal singing conversational first-person lyrics, small and honest, slight tape wobble, short bridge
- Keepers (release-candidate quality)
- 2 of 10
- Usable with editing
- 5 of 10
- Failures
- 3 of 10
- Belted final chorus despite exclusion
- 2 of 10
- Lyric drift into abstraction
- 4 of 10
Small Hours
96 BPM · G maj · 4:29Run 2, the strongest take, later rewritten and re-sung for release as Small Hours.
“Publishing the bad takes was the point. Taste is a public act now.”
The practical lesson: the recipe's stated reliability of 4 out of 5 held up, but only if 'usable' is your bar. If your bar is 'release candidate', budget five runs per keeper and write your first lyric line yourself. Every run where June supplied a concrete opening image ('the lamp is still on') scored higher on lyric quality than the runs that started blank.
Independent educational resource. Not affiliated with or endorsed by Suno, Inc. Suno is a trademark of Suno, Inc.