Sensif.aiAI signal desk

VIEWS

Today Stories Radar Briefings Listen Watchlist

Loading…

Today Stories Radar Briefings Listen

DESK

Frontier AI Research Countries Open vs Closed Compute What moved

Loading…

Radar · Phonon 31m 31m | Sensif.ai

Radar · Release

Phonon 31m 31m

Gradium · released Oct 22, 2026

Open-weight (permissive)text-to-speechtextaudio

Openness taxonomy

access, licence, and disclosure — never a single open/closed binary

Weights availability: Model released under Apache 2.0; fits in 64.5 MB
Licence: Apache 2.0
Licence category: permissive
Commercial use allowed: Yes
Derivatives allowed: Yes
Redistribution allowed: Yes
Field-of-use restrictions: unknown
Training code availability: unknown
Inference code availability: unknown
Architecture disclosure: unknown
Training data disclosure: unknown

openness confidence

80

Source states English Phonon 31m ships under Apache 2.0
Source describes Phonon as open-sourced

Deployability

raw values as extracted — free text, never inferred from branding

Self-hosting: On-device deployment
On-premise: unknown
Edge deployability: Designed for on-device use; reported to fit in 64.5 MB
API availability: unknown
Regions available: unknown

Benchmarks

as reported — vendor-supplied unless otherwise noted

No benchmark data on this release yet.

National-security pathway & evidence

from the assessed source item, when one is linked

Impact: lowConfidence: mediumNat-sec: noneindependent reporting

Evidence limitations: The item provides a single report of Gradium's release and performance figures; no independent reproduction is included.

Gradium Open-Sources Phonon, a 31M-Parameter On-Device TTS Model Under Apache 2.0

Linked research papers

from paper_model_links

No linked research papers yet.