Gradium open-sources Phonon: 31M-param on-device TTS with 0.02s latency, Apache 2.0
natesiggard · x · 2026-10-24
GradiumAI open-sourced Phonon, a suite of on-device TTS models under 50M parameters targeting naturalness and accuracy across languages, with Phonon-31m under Apache 2.0. Benchmarks show warm-median time-to-first-byte of 0.02s and 0.6s load — far ahead of Kokoro (3.2s/0.08s) and Qwen3-TTS (3.0s/0.12s). It's now supported in pipecat-local-tts-server v0.6.0; the required mlx-audio bump also sped up other models by 14–20%.
More from Infra
- Tencent's Hy-Nomas world model needs at least 24GB VRAM, dev complains — cocktailpeanut · 2026-10-24
- Humansand Speeds Up NVFP4 Decode Kernel by Up to 10% While Improving Accuracy — noahdgoodman · 2026-10-24
- Recursive's 40B Flint Hits ~200 tok/s on M3/M5 Ultra via MLX-VLM with Fused Metal Kernels — timshi_ai · 2026-10-24
- LiteLLM breach lesson: AI gateway logs are a secrets store — alvelda · 2026-10-24
- US electricity prices up 34% at investor-owned utilities since 2020 vs 19% public — randal_olson · 2026-10-24
- ExecuTorch now builds on Windows Snapdragon X Elite, demoing on-device Llama 3.2 1B — carrycooldude · 2026-10-24