Edge0 releases compact TTS model Audio8-TTS-Realistic-0.6b with...

Edge0 releases compact TTS model Audio8-TTS-Realistic-0.6b with Russian language support

Edge0 has released an open, compact TTS model, Audio8-TTS-Realistic-0.6b, with 601,159,424 parameters (excluding the codec). It improves intonation and expressiveness compared to the Preview version, clones voices from a recording and its text without fine-tuning, and supports 23 languages, including Russian.

Compiled by Sergey KostenchukPublished 2026-10-30Updated 2026-10-30

2026-10-30 Research OpenAI

🎙 Edge0 releases compact TTS model Audio8-TTS-Realistic-0.6b with Russian language support

The 601-million-parameter model improves intonation and expressiveness compared to the Preview version, clones voices from short recordings and their text without fine-tuning, and speaks 23 languages—12 of which are new, including Russian.

🌍 The open 0.6B TTS model, featuring a 44.1 kHz codec, DualAR architecture, and SGLang Omni adapter (OpenAI-compatible API /v1/audio/speech), provides a ready-to-use stack for local and server-side speech synthesis. In a developer-provided demo comparison, it competes with systems 3–7 times larger, though this is a comparison by the developer rather than a blind test.

👤 Weights, code, and demos are available on Hugging Face: the model can be run on a GPU or tried in a browser. Russian is among the new languages, but quality varies by language—pronunciation errors are possible.

Source 1: https://huggingface.co/Edge0/Audio8-TTS-Realistic-0.6b

Source 2: https://edge0-ai.github.io/Audio8_TTS/realistic-0.6b/

Sources