The tool doesn't have to be a distinct model. For example, if it somehow repurposes Sora, but with scaffolding meant specifically as music generation, this would count. The tool must be first-party. Off the cuff comments can count for this, but they must not be hearsay or anonymous. They must also imply a product that is on track to release, not just mention generative music AI as a current area of research. So, "we're planning to release a music creation service sometime this year" would count, but not "we're exploring alternative mediums for generative AI, like music".
Market closes at July 1st, 00:00 UTC.
Update 2025-10-26 (PST) (AI summary of creator comment): The tool must have audio modality (i.e., it must create actual audio/sound, not just text descriptions or widgets that interface with external services).
1,000People are also trading
Just placed a light bet of 100 against on this. My reasoning is that OpenAI has been scared of the music industry for at least 2 years. And, I don't see that changing soon.
OpenAI let Midjourney get a large business on taking more risk than them on image generation. I think they will let the audio generation companies take the lead there as well.