Stable Audio 3.0 Medium

model·generative audio model·active

Stable Audio 3 exact tier for higher-musicality long-form generation.

Verified facts

Official site	https://huggingface.co/stabilityai/stable-audio-3-medium ↗
Geography	unknown
model id	stable-audio-3-medium
parameter count	2B
api	unknown
inputs	text; optional audio
license	Stability AI Community License
outputs	44.1 kHz stereo music and sound
weights	public
modalities	44.1 kHz stereo music and sound; optional audio; text
dependencies	SAME semantic-acoustic autoencoder
model family	Stable Audio 3.0
access method	downloadable weights; self-hosting
exact version	stable-audio-3-medium
repository links	https://huggingface.co/stabilityai/stable-audio-3-medium
evaluation results	unknown
evidence separation	developer claims: Stability AI claims licensed training data and sub-two-second generation on specified hardware. · measured results: · independent evaluations:
developer maintainer	Stability AI
material us relevance	Publicly accessible in or materially offered to the United States market as of snapshot.
documented capabilities	higher-musicality long-form generation; maximum documented duration: up to 6 minutes 20 seconds; variable-length generation
products exposing model	StableAudio.com; Stability AI API
commercial use conditions	Community License below $1M annual revenue; Enterprise license above threshold.
known geographic or policy restrictions	Community License revenue threshold applies to open weights
safety watermarking disclosure provenance	Uploaded audio is subject to rights restrictions and moderation controls