AI just learned to create six-minute songs — here's how

Stability AI has released Stable Audio 3.0, a new family of audio models that can generate songs up to six minutes long using licensed training data. As reported by TechCrunch, this extends the length of music that AI models can generate, with the company emphasising its use of properly licensed content for training.

What makes Stable Audio 3.0 different?

The standout feature is length — six minutes represents a substantial increase over previous AI audio models that typically generated clips of 30 seconds to two minutes. This duration allows for complete song structures with verses, choruses, and bridges rather than short loops.

Equally important is Stability AI's approach to training data. The company explicitly states these models were trained on licensed content, addressing one of the biggest controversies in generative AI. This positions Stable Audio 3.0 as a more ethically sound option for commercial use.

The release comes as the UAE leads global AI adoption, with local creative industries increasingly exploring AI tools for content production.

How does this impact music creation?

For content creators, the six-minute capability means generating complete background tracks for videos, podcasts, or commercial projects without needing multiple clips. Musicians can create full demos or explore ideas without traditional recording equipment.

The licensed training data also means creators can potentially use these tracks commercially without the legal uncertainty that surrounds other AI-generated content. Copyright clearance is essential for many professional applications.

However, the quality and creativity of six-minute generations remain to be tested in real-world scenarios. Extended AI audio often struggles with maintaining coherence and avoiding repetitive patterns.

What's the broader context?

Stability AI's move reflects the industry's push towards longer, more sophisticated AI-generated content. show major companies integrating AI across creative workflows.

The emphasis on licensed data addresses ongoing concerns about AI training practices. As AI tools face increasing scrutiny over training practices, companies that can demonstrate proper licensing may gain competitive advantages in commercial markets.

For the UAE's growing tech sector, tools like Stable Audio 3.0 offer opportunities for local content creators and businesses to explore AI-assisted production workflows.

Availability and access

Stability AI has not yet announced pricing details or specific availability dates for Stable Audio 3.0. The company describes it as a 'family of audio models', suggesting multiple tiers or specialised versions may be available.

Access methods and subscription options remain unclear, with more details expected in the coming weeks. Previous Stable Audio versions offered both free and paid tiers with different usage limits and capabilities.

Frequently Asked Questions

What is Stable Audio 3.0?

Stable Audio 3.0 is a new family of AI audio models released by Stability AI, capable of generating music and audio tracks. The top model can produce songs up to six minutes long.

How long can Stable Audio 3.0 generate songs?

The top model in the Stable Audio 3.0 family can generate songs up to six minutes in length, significantly longer than previous AI audio tools.

Was Stable Audio 3.0 trained on licensed data?

Yes, Stability AI states that all Stable Audio 3.0 models were trained on licensed data, addressing copyright concerns common with other AI music generators.

When will Stable Audio 3.0 be available?

Stability AI has not announced specific availability dates or pricing for Stable Audio 3.0. More details are expected in the coming weeks.

Written by

Share

Copied