Gradium open-sources on-device TTS model Phonon with 1.7% WER, beating Kokoro
amos_gyamfi · x · 2026-10-27
Gradium released Phonon, its first open-weight, on-device text-to-speech model designed to be small enough to ship in any app, with private offline operation on iOS and macOS. The author also shares an Ollama-based deployment path and benchmarks it against Kokoro TTS and PocketTTS for local voice generation.
Official numbers (tested on 500 sentences other models get wrong — phone numbers, emails, postal codes):
- Phonon 31m (English): 1.70% WER
- Phonon 40m multilingual: 2.11%
- Phonon 40m: 2.22%
- Kokoro: 3.50%
- Kitten TTS: 7.40%
Alongside the release, Gradium extended its seed funding to $100 million with NVIDIA joining as an investor and opened a Silicon Valley office to scale real-time voice AI.
Related event: Gradium Open-Sources On-Device TTS Model Phonon(2 posts)→
More from Infra
- Nvidia and Micron drove a third of S&P 500 profit growth last quarter — LexSokolin · 2026-10-27
- WSJ: Broadcom CEO sealed Anthropic's $63B chip debt deal with loan guarantees and higher rates — rohanpaul_ai · 2026-10-27
- Digital Sovereignty Is Now a Cloud Architecture Decision, and AI Pipelines Make It Harder — DavidLinthicum · 2026-10-27
- Houmo LQ50-24 M.2 accelerator claims 160 TOPS but real-world tests are scarce — lazy-kozak · 2026-10-27
- TRL's tree attention adds linear attention support with up to 3.46x speedup — SergioPaniego · 2026-10-27
- Tested: With Qwen3 27B and Dflash2, a Q4 Draft Model Gives the Biggest Speedup — milpster · 2026-10-27