Udio lets you make 15-minute tracks

If you have tried generating songs with AI, you know the frustration of a model forgetting its own chorus thirty seconds later.

Udio just rolled out an update targeting that exact limitation. The platform expanded its maximum track length to 15 minutes, giving users room for sprawling progressive epics and extended trance mixes.

More importantly, Udio boosted the context length for audio extensions from 30 seconds to 2 minutes.

Remembering the melody

This means the model can now reference material from much earlier in the track while creating an extension.

In practice, creators can repeat motifs and achieve far more consistent choruses across multiple sections instead of watching the song drift off.

Why it matters

For music creators, context is everything. AI audio generation has always struggled with short-term memory, which made building authentic verse-chorus pop arrangements nearly impossible without manual editing.

Expanding context to 2 minutes tackles the hardest part of generative music: structural coherence over time.

While 15-minute tracks sound impressive, we think the real test will be whether the model can maintain melodic focus without turning the later minutes into repetitive mush.