Udio introduces new udio-130 music generation model and more advanced features

Brain Titan
2 min readMay 30, 2024


Udio introduces a new udio-130 music generation model that generates 2 minutes of audio to help create more coherent and structured tracks.

This means that music generated using this model will not appear disorganized, but will have a clear beginning, middle, and end, just like a complete song. As a result, it is much easier to produce music using this model to get a logical and structured piece of music, rather than some random pieces of sound thrown together.

New advanced controls have also been added: random seed settings, cue word or lyric intensity adjustment, and more.

- Random Seed Setting: Allows the user to use the same random seed when generating music clips, thus making the generated clips repeatable. Using the same seed allows certain specific characteristics to be maintained each time it is generated, even if the cue word or lyrics change.

- Cue or Lyrics Intensity Adjustment: Allows the user to control how much the cue or lyrics affect the generated music. Increasing the cue word intensity enhances the match between the music and the cue words, but may make the music sound less natural. Conversely, decreasing the intensity of the lyrics can make the vocals more natural, but may sometimes cause the lyrics to be ignored.

- Clip start time control: This feature allows the user to specify the starting position of the generated music clip in the whole song. For example, 0% means to start generating from the beginning of the song, 50% means to start from the middle, and 90% means to generate from the end part of the song. This is especially useful when used in conjunction with extensions to help the user generate music from a specific part of the song.

- Generation quality slider: This slider allows the user to trade-off between generation speed and music quality. Users can choose a higher generation speed to explore different musical ideas faster, or a higher quality for better musical results.

experimental feature is only available to Pro subscribers and will be rolled out more broadly in the coming weeks.