Open-source TTS: Phonon models under 50M params run on-device with 7ms latency
alex_verem · x · 2026-11-05
Gradium AI has open-sourced Phonon, a suite of on-device TTS models all under 50M parameters, with Phonon-31m released under Apache 2.0. The team claims new standards for naturalness and accuracy across languages plus the broadest inference support to date. The pitch: fully local voice, first audio in 7ms, no data leaving your machine.
More from Models
- Claude trains a model instead of writing code: 1257-param pixel-art generator in 10 minutes — victormustar · 2026-11-05
- MoE beats dense in matched OOD finetune: DeepSeek V3.1 hits 0.804 CE vs Llama 405B — kalomaze · 2026-11-05
- Grok Bot tests sharing disclosure: files, skills and credentials visible before you share — XFreeze · 2026-11-05
- Ari agent post-trains Qwen3.5-9B unsupervised for 16h, reward 0.118→0.629 — ypatil125 · 2026-11-05
- Happenstance scores 84.7 on PeopleSearchBench's 119 queries, nearly 20 points above next best — alex_teichman · 2026-11-05
- Google Research unveils Agile Deliberation, an LLM framework for teaching VLMs subjective categories — RanjayKrishna · 2026-11-05