NVIDIA Fugatto 1

model·generative audio model·unknown

Research model for free-form audio synthesis and transformation from text and optional audio.

Verified facts

Official site	https://research.nvidia.com/publication/2025-04_fugatto-1-foundational-generative-audio-transformer-opus-1 ↗
Geography	unknown
api	unknown
inputs	text; optional audio
license	not publicly disclosed
outputs	audio; music; singing voice
weights	not_released
modalities	audio; music; optional audio; singing voice; text
dependencies	unknown
model family	Fugatto
access method	research demonstrations
exact version	Opus 1
repository links	unknown
evaluation results	unknown
evidence separation	developer claims: NVIDIA describes Fugatto as a versatile instruction-following audio model. · measured results: · independent evaluations:
developer maintainer	NVIDIA
material us relevance	Publicly accessible in or materially offered to the United States market as of snapshot.
documented capabilities	audio synthesis; audio transformation; instruction composition and negation
products exposing model	unknown
commercial use conditions	No public access or commercial-use terms verified.
known geographic or policy restrictions	Research-only page; no public weights or hosted API verified
safety watermarking disclosure provenance	unknown