ElevenLabs’s new music generation model can switch genres mid-track

TechCrunch reports that ElevenLabs has launched a new music generation model whose key capability is allowing users, after a song has been generated, to regenerate one specific section without affecting the rest of the song. For music creators, video editors, and advertising and social content producers, the significance lies in reducing the cost of repeatedly revising AI music: in the past, if you were dissatisfied with a particular bridge, transition, or chorus, the common workflow might be to regenerate the entire song and then pick usable fragments from multiple versions; the new model instead attempts to shrink the editing granularity to local parts of the song, letting users keep the melody, rhythm, or arrangement they are already happy with and replace only the section that needs changing. The title also notes that this model can let a song switch genres midway, meaning it is not just doing ordinary audio inpainting but may support more creative structural changes, such as transitioning from one style to another. However, the information provided in the original article is very limited, with no mention of the model name, official launch date, pricing, licensing terms, available regions, output audio quality, whether commercial use is supported, or how it integrates with existing ElevenLabs audio products. Therefore, this news is currently best regarded as a product update on ElevenLabs' generative music editing capabilities, rather than a complete technical release or an evaluable model report.