Curiosity · AI Model
Udio v1.5
Udio v1.5 is the July 2024 update to Udio's text-to-music model, built by a team that spun out of Google DeepMind. The model produces high-fidelity music with vocals and instrumentation from a text prompt plus optional lyrics, extending generations to 15-minute tracks and improving acoustic realism, stereo mixing, and genre coverage. Udio exposes inpainting, extension, and remix tools, and is typically rated alongside Suno as the other leading music-gen service.
Model specs
- Vendor
- Udio
- Family
- Udio
- Released
- 2024-07
- Context window
- 1 tokens
- Modalities
- text, audio
- Input price
- n/a
- Output price
- n/a
- Pricing as of
- 2026-04-20
Strengths
- Industry-best acoustic realism for many genres
- Long-form generation up to 15 minutes
- Inpainting and extension for iterative editing
- Detailed genre and mood tagging
Limitations
- Closed — credit-based subscription
- Vocals occasionally less intelligible than Suno's
- Copyright and training-data litigation pending
- Master-quality production still requires DAW polish
Use cases
- Long-form music drafting
- Acoustic and instrumental composition
- Remixing / inpainting / extension of existing outputs
- Songwriting with detailed lyric control
Benchmarks
| Benchmark | Score | As of |
|---|---|---|
| Max track length | ≈15 minutes | 2024-07 |
| User-preference vs Suno v3.5 (LMSYS-style) | roughly tied, slight Udio lead on acoustic | 2024-12 |
Frequently asked questions
What is Udio v1.5?
Udio v1.5 is Udio's July 2024 music-generation model, producing up to 15-minute tracks with vocals and strong acoustic realism.
Who built Udio?
A team of researchers that spun out of Google DeepMind, with backers including a16z and Will.i.am.
Is Udio commercially usable?
Paid subscribers have commercial usage rights under Udio's terms, subject to the ongoing legal landscape around AI-generated music.
Sources
- Udio product site — accessed 2026-04-20
- Udio v1.5 release note — accessed 2026-04-20