Particle.news
Download on the App Store

OpenAI Reportedly Building AI Music Generator With Juilliard-Backed Training Data

Key decisions on product format, pricing, release timing remain undecided.

Overview

  • Multiple outlets citing The Information report that OpenAI is developing a model to create full songs from text or audio prompts.
  • Reported capabilities include adding music to existing videos and generating instrument accompaniments such as guitar tracks for recorded vocals.
  • OpenAI is collaborating with Juilliard students who are annotating musical scores to provide high-quality training data.
  • It remains unclear whether the tool will launch as a standalone product or be integrated into ChatGPT or Sora, and no release date or pricing has been announced.
  • The effort builds on OpenAI’s MuseNet and Jukebox work and would compete with tools from Google and Suno as the company broadens its creative AI lineup following Sora’s rollout.