Project AURIONAI Use-Cases Research & Innovation
music and audio

Stable Audio 3 Medium

Stability AI open-weight audio model (also Small Music and Small SFX), trained on licensed AudioSparx and CC Freesound data.

Stability AIStability AIUKStable AudioFamily
Stability AI
Levelmodel
Classmusic and audio
Parameters2B
Contextnot in the notes
Released18 May 2026
Opennessopen weights restricted
Runs onconsumer GPU
Statusactive

Benchmark Results

independent
Artificial Analysis Music Arena v1.1generative media
secondary capture of AA postindependent
998
8001,0001,200

Score in Elo; the axis starts at 800, as Elo has no zero. Higher is better.

AA Music Arena v1.1, InstrumentalOct 2026 alphasignal.ai, digg.com
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

Lineage

successor
Stable Audio 2.5 to Stable Audio 3 Medium: successor, medium confidenceStable Audio Open 1.0 to Stable Audio 3 Medium: successor, low confidenceStable Audio 2.5Stable Audio 2.5Stability AIStable Audio Open 1.0Stable Audio Open 1.0Stability AIStable Audio 3 MediumStable Audio 3 MediumStability AI, May 2026
Stability AIStable Audio 2.5 is the parent (successor)medium confidencenotes
Stability AIStable Audio Open 1.0 is the parent (successor)low confidencenotes
Where this comes from
Parent and child relations from model cards, Hugging Face base_model fields and the genealogy tables of the research notes. The tree shows two generations each way, 3 models. Confidence and evidence are the notes' own.

R&D Topics

Text-to-music and song generationproduced by

Identity

Developer
Stability AI
Family
Stable Audio
Released
18 May 2026 sourcesregistry date
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
diffusion Transformer latent diffusion, T5Gemma conditioning, adversarial post-training
Total parameters
2B sources
Active parameters
–
Checkpoint parameters
–
Context length
not in the notes
Input
text
Output
audio
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
open weights restricted
Licence
Stability AI Community Licence sources
Access routes
open weights download
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
consumer GPU
Footprint
not in the notes
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
English only
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Industry Use Cases

No industry use case links this model.

Sources

  1. huggingface.co/stabilityai/stable-audio-3-medium https://huggingface.co/stabilityai/stable-audio-3-medium
  2. Text-to-Audio Models – Hugging Face huggingface.co https://huggingface.co/models?pipeline_tag=text-to-audio&sort=trending
  3. Stability AI - Wikipedia en.wikipedia.org https://en.wikipedia.org/wiki/Stability_AI

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Stable Audio 3 Medium", https://aurion.semai.id/models/stability-ai-stable-audio-3-medium/, records of October 2026.