DeepSeek's MoE model with 552B total parameters and 8-16B active per token. 1M-token context, MIT license: free commercial use. Built for online serving, too large for local use on a single PC.
Phonon
Released on Oct 24, 2026
Compact TTS model family from Gradium, in Ollama: one command to speak, an OpenAI-compatible endpoint and streaming read-aloud of replies. Apache 2.0 in English, CC BY-NC 4.0 in the 40M multilingual variants.
Features
- Streaming TTS in locale
- Endpoint OpenAI-compatibile
- MLX su Apple Silicon
- 40M: ~60 MB GGUF
Version history
- Phonon (famiglia)Latest
Famiglia TTS compatta di Gradium: phonon (inglese, Apache 2.0) e varianti 40M per fr/de/es/pt/multilingue (CC BY-NC 4.0, GGUF ~60 MB q4k). In Ollama con accelerazione MLX, 35% più veloce del runtime di riferimento su M5 Max; endpoint /v1/audio/speech OpenAI-compatibile, 24 kHz mono.
Tags
Useful links
Discover Phonon
Visit the official site to download, learn more and start using Phonon.
Related resources — Models
The most advanced open image model: excellent prompt adherence, solid anatomy, full fine-tuning. The standard for serious work.
Very few steps, high quality, and strong text rendering. The distilled FLUX variant, built for speed.

