OpenAI is reportedly developing a new generative music tool capable of creating compositions from text and audio prompts, according to The Information. The project would allow users to generate original music or add instrumentals — such as guitar accompaniments — to existing tracks, expanding the company’s growing portfolio of creative AI products.
While details remain limited, sources familiar with the project say OpenAI has been exploring ways to apply its multimodal AI systems to music production. The tool could eventually be used to score videos automatically or generate background music for user-created content. It’s unclear whether OpenAI plans to release the model as a standalone product or integrate it into existing platforms such as ChatGPT or the video-generation system Sora.
One source cited by The Information said the company is collaborating with students from the Juilliard School to annotate musical scores, providing structured training data for the model. That approach suggests OpenAI is focusing on quality-controlled datasets to help the system learn musical structure, rhythm, and instrumentation — areas that have historically been challenging for AI models to master.
OpenAI has experimented with generative music before, though its earlier efforts came before ChatGPT’s release and were never widely deployed. In recent years, the company’s attention has shifted toward audio models that handle speech synthesis and transcription, including tools for realistic text-to-speech generation. The new music project would mark a return to generative creativity, positioning OpenAI to compete more directly with platforms such as Google’s MusicLM and Suno, both of which already allow users to create songs from text prompts.
If launched, OpenAI’s music generator could further blur the lines between creative software and AI artistry, raising new questions around authorship, licensing, and the role of musicians in AI-assisted production. It would also reflect a broader trend among major AI developers to expand into music — one of the few creative domains still relatively untouched by mainstream generative tools.
