Hugging Face releases on-device TTS model Phonon
What happened
On 5 November 2026, Hugging Face published a retrospective on its blog about the release of Phonon, an on-device speech synthesis model. According to the post, Phonon is a small TTS model that runs offline on local devices, comes in roughly 65MB and 75MB versions, supports five languages (English, French, German, Spanish and Portuguese), includes more than 100 voices, and is said to have topped a related leaderboard. The post does not disclose an earlier release timeline, the name of the leaderboard or evaluation details, and gives no comparison data with other models. The story remains at the stage of an official release retrospective, with no third-party evaluation or follow-up version updates yet.
Written by AI from the reports · updated 3 h ago
Timeline
Follow the reports to see the event from each side.
- Hugging Face BlogSelectedA Closer Look at Phonon TTS. Launching High-Quality On-Device Speech That Tops the Benchmarks.
Two weeks ago, we opened up Phonon, our small text-to-speech model that runs on your own device. It needs no server and no network, and what you type stays with you.
Heat of this event
Not enough continuous observations yet to draw a trend.