Jukebox
OpenAI open-sources Jukebox: a neural net that generates music and singing from scratch. Here's why audio generation matters for the AI economy.
Why it matters
OpenAI expands beyond text into generative audio with an open-source model capable of multi-genre music and vocal synthesis. This marks a significant capability expansion and signals the trajectory toward multimodal foundation models—a strategic move that influenced the entire industry's approach to content generation.
The key facts
5 to knowOpen-source model release with weights and code
Capability: music generation including rudimentary singing
Raw audio generation across multiple genres and artist styles
Published April 30, 2020 — early exploration phase of generative audio
Historical significance: precursor to later multimodal model development
Go to the source
OpenAI Blogopenai.com
Publisher excerpt: We’re introducing Jukebox, a neural net that generates music, including rudimentary singing, as raw audio in a variety of genres and artist styles. We’re releasing the model weights and code, along with a tool to explore the generated samples.